OpenAI, Google Fight for Dominance as Consumer AI Retains Power Users

消费级AI锁定核心用户，OpenAI与Google竞逐霸主地位 消费级AI格局正整合为一个赢家通吃型市场，头部实验室拥有巨大的分发优势；然而算力瓶颈以及企业倾向于稳健、渐进式功能的偏好，为敏捷的初创公司留出了宝贵的窗口期，使其能够通过具有鲜明主张、针对特定任务的界面来吸引核心用户。 本次对话概述了一个高度集中的消费级AI市场，其中OpenAI的ChatGPT占据显著领先地位，但Google的Gemini正凭借Veo和Nano Banana Pro等前沿创意模型在桌面端获得强劲增长势头。投资者评估了各主要产品发布喜忧参半的表现，指出尽管像Perplexity Comet浏览器和Claude工作流功能这类技术型生产力工具在专业消费者群体中蓬勃发展，但备受期待的社交实验（如Sora 2）作为独立网络却难以实现用户留存。展望2026年，专家组预测将出现大规模模板驱动的多模态和定制应用生成浪潮，并强调敏捷初创公司迎来了一个独特的机遇窗口，能够通过推出具有鲜明主张、针对特定任务的用户界面，超越受算力瓶颈制约的科技巨头。 消费级AI格局日益呈现出赢家通吃的态势，其特征是极高的用户忠诚度，仅有9%的消费者订阅多种AI工具。 (0:02) ChatGPT以8亿至9亿的周活跃用户数保持市场主导地位，尽管其同比增长放缓至23%，而Gemini的桌面端用户基数则激增155%。 (0:02) OpenAI和Google在2025年采取了截然相反的生态策略：OpenAI将功能整合进核心ChatGPT界面中，而Google则大力倾向于推出独立产品和定制UI。 (2:23) 图像与视频生成模型已从侧重评估艺术美感，转向优先考量深度物理推理、逻辑一致性以及空间准确性。 (4:04) 尽管进展迅速，但前沿图像模型在处理复杂的多步骤空间推理和上下文文本排版时仍面临困难，这在“大富翁测试”（Monopoly Test）的失败中得到了体现。 (5:39)

According to an a16z analysis, consumer AI at the end of 2025 is trending toward a winner-take-most market anchored by ChatGPT's default distribution and enterprise spillover, even as foundation labs struggle to design successful standalone consumer products. This organizational friction leaves substantial openings for opinionated startups and specialized prosumer tools to capture high-value workflows. Incumbent Market Concentration and Enterprise Spillover (0:02) Early market data indicates that general conversational AI is consolidating into a winner-take-all dynamic. ChatGPT leads with an estimated 800 to 900 million weekly active users, insulated by high consumer inertia: only 9% of consumers pay for more than one AI subscription, and fewer than 10% of ChatGPT users visited competing LLM providers for most of the year. While ChatGPT's year-over-year growth slowed to 23%, Google's Gemini is expanding rapidly on desktop at 155% year-over-year, reaching 50% of ChatGPT's footprint on Android but only 17% on iOS. The speakers emphasize that enterprise adoption serves as an underappreciated moat for consumer retention. OpenAI reported enterprise usage growing roughly 8x to 9x year-over-year. Because enterprise contracts mandate ChatGPT in professional environments, it naturally cements itself as the default assistant for personal tasks, reinforcing ChatGPT's status as the household brand name of AI. Specialized Niches for Technical and Prosumer Workflows (10:21) Rather than unseating general chat leaders, competing platforms are finding traction by specializing in narrow user profiles. Anthropic has captured hyper-technical developers through capabilities like artifacts, the Model Context Protocol (MCP), and Claude Code. However, complex interfaces hinder Anthropic's mainstream consumer penetration, highlighted by survey data indicating that three times more US teenagers have used Character AI than Claude. Conversely, Perplexity has established momentum within the prosumer productivity category. Its agentic Comet browser sustained higher launch traffic than ChatGPT's Atlas browser despite Perplexity's smaller distribution footprint. Bolstered by targeted acquisitions of agentic startups and email assistant features, Perplexity is leveraging repeatable browser-level automations to retain power users. Structural Obstacles in AI Social and Standalone Interfaces (16:28) The discussion highlights that major labs consistently fail to build opinionated, standalone consumer products, citing low adoption for initiatives like OpenAI's Pulse, Atlas, and group chats, alongside Google's Gems and Opal—with Google's NotebookLM noted as a rare exception. Under the speakers' 'inception theory,' products fail when their design conflicts with user intent: ChatGPT is fundamentally a productivity tool driven by utility ('help me'), whereas social platforms succeed on entertainment ('entertain me') and personal connection ('see me'). This dynamic explains why dedicated creative applications like Sora 2 struggle as standalone social networks. Because pure AI generation removes the human vulnerability central to social status games, power creators treat Sora 2 as a creation utility analogous to CapCut, immediately exporting generated videos to platforms like TikTok, Instagram, and X rather than consuming content natively within the app. Multimodal Evolution and Startup Monetization Advantages (33:25) Creative generation is transitioning from aesthetic realism to deep reasoning, search integration, and multi-input workflows. Although models can now maintain spatial coherence in video and research accurate infographics via search, multi-step spatial reasoning remains limited, as demonstrated by failures on the 'Monopoly test' of replacing board elements sequentially without overlap. Despite these model limitations, application startups maintain key advantages over major labs. Incumbent product managers face institutional incentives to make safe, incremental bets, while foundation labs face severe internal compute constraints, forcing them to balance model training against inference demands. Unburdened by training infrastructure, startups like Krea, Gamma, and Wabi focus entirely on user experience and hybrid consumption pricing, driving consumer AI net revenue retention past 100% among highly monetizable power users. While major labs will continue expanding core models into integrated enterprise app ecosystems and unified multimodal architectures, the source concludes that building opinionated consumer interfaces remains outside their primary competency. The central unresolved challenge heading into 2026 is whether specialized startups can maintain their interface advantage as foundational models increasingly absorb image, video, and reasoning into single systems.

a16z对2025年底消费级AI格局的研判指出，通用LLM赛道呈现出极强的“赢家通吃”特征，ChatGPT确立了默认品牌壁垒；然而，大模型实验室在打造具有鲜明观点的消费端独立界面与社交产品上普遍受挫，为专注于多模态工作流、专业消费者界面的初创公司留出了破局窗口。 赢家通吃格局与巨头的产品形态分歧 (0:02) 2025年消费级AI市场展现出显著的垄断趋势。数据显示，仅有9%的消费者同时付费订阅ChatGPT、Gemini、Claude和Cursor中的两款及以上；全年中不到10%的ChatGPT用户会访问Gemini等竞品。目前ChatGPT以每周约8亿至9亿活跃用户占据主导，虽然其同比增速放缓至23%，但用户使用频率高达平均每周25次（每日3至4次）。与此同时，Gemini在桌面端依靠模型发布实现155%的年同比增长，并在移动端达ChatGPT的40%（Android端达50%，iOS端仅17%）。Claude、Grok与Perplexity的使用规模则各自仅相当于ChatGPT的8%至10%。 在落地形态上，OpenAI与Google采取了不同策略。OpenAI力求平台整合，将投票、群聊、购物、研究等功能直接汇聚于ChatGPT主界面，仅有Sora作为独立视频应用推出；相比之下，Google更倾向于分散发布，通过Google AI Studio、Google Labs及现有全家桶生态推出独立工具和网站，探索超越传统单一对话框的界面形态。 多模态演进：从单一画质转向深层推理与搜索结合 (4:04) 图像与视频生成在过去18个月内发生了根本演进，行业重心由Midjourney早期的艺术美感转向逼真的物理规律与深层推理。现在的模型已能保持时空连贯性，例如视频背景中的行人和车辆能按正确方向平滑移动而不产生畸变。像Nano Banana Pro等模型支持同时输入多张图像与提示词，并借助搜索集成获取真实世界参考数据，大幅降低虚构细节的概率，可生成包含准确公司标志的信息图表。 然而，图像模型的复杂多步骤推理仍存瓶颈。在“大富翁棋盘测试”中，要求模型抹去原有地产名并替换为AI机构与初创公司名称，即便顶尖的GPT Image 1.5也难以在不产生重叠、重复或遗漏的前提下同时完成文本擦除、情境构思与精确空间排布。不过，ChatGPT图像模型在角色一致性（即跨轮次生成中保持角色外貌与风格恒定）上取得显著突破，为分镜脚本设计带来了实用价值。 AI社交探索受挫与专业消费者工具的崛起 (16:28) 对话者基于“盗梦理论”（即剥离产品表层探究底层核心情感驱动力）分析认为，ChatGPT的核心诉求是“帮我提高效率”，而TikTok和Instagram满足的是“娱乐我”与“排解孤独/寻求关注”。将生产力工具强行植入社交场景难度极高，ChatGPT群聊功能最终也仅停留在行程规划等实用层面，未能促成深度人际连接。Sora 2的轨迹同样印证了这点：它虽被设计为带有朋友客串功能的类TikTok信息流，但现实中极少有用户在应用内消费与互动，而是被创作者当作生成病毒式内容的素材制作工具，再转录至TikTok、Instagram等外部平台，其真实竞品更接近CapCut而非社交网络。 相比之下，面向专业消费者（Prosumer）的工作流工具表现亮眼。Perplexity推出的Comet浏览器在发布初期流量与持续活跃度上甚至超越了ChatGPT的Atlas浏览器界面；其通过集成智能体模型与自动化定时工作流，成功留存了高阶用户。类似Gamma（文档一键转演示文稿）、Granola（保留历史上下文的会议纪要）以及Cursor等工具，正在通过解决具体的复杂任务捕获高黏性用户。 垂直壁垒与初创公司的战略机遇 (31:54) 主流AI实验室虽然擅长迭代底层模型，但打造独立、风格鲜明的消费级UI并非其核心能力。Google与OpenAI推出的诸多消费级实验性功能（如Pulse、Atlas、Sora、Stitch、Gems等）多未形成持久黏性，仅Google的NotebookLM被视为少数例外。与此同时，大型实验室受困于算力分配瓶颈：它们必须在下一代模型预训练与日常推理、娱乐性流量与代码智能之间权衡算力分配；此外大公司中层产品经理出于晋升与合规考量更倾向于平庸的微调，导致产品设计缺乏鲜明主张。 这为应用层初创公司带来了结构性红利。初创企业无需承担算力分配冲突，能够专注于特定垂直领域。在商业化层面，AI消费级软件正展现出前所未有的财务指标，依托“订阅保底+超额Token计费”的混合模式，部分产品实现了原本专属于企业级SaaS的100%以上净收入留存率（NRR），高价值核心用户正在撑起整个商业闭环。 总体而言，节目认为2025年底的消费级AI正处于从纯文本对话向全模态工作流迁徙的节点。尽管巨头凭借分发优势牢牢把持着基础大模型入口与企业默认工作流，但在具体垂直场景中，能够设定合理约束、打造鲜明特色界面的应用层初创公司依然拥有广阔的突破空间。

[31:54] None of those are really working, and I think it's because it's not the core competency of these companies anymore to build opinionated standalone consumer UI. [16:28] I want my dad to love me [16:28] help me be better, more productive, and more efficient. [16:28] entertain me [16:28] I'm lonely, I want to be seen and connected. [14:04] Evan Spiegel was chief product officer at Meta [14:04] I don't know what to do. [22:11] the best product across all of them at doing that kind of complex work. [22:11] dumb it down even more [23:37] find the kid in the red t-shirt. [23:37] it basically makes it seem like you're a native speaker [25:09] the steepest slope [39:31] editing every little pixel in your Google Slides to get it to fit. [38:42] as a sort of semi-normal person. [41:37] What is a most unhinged thing I said this year?

[0:02] For most of the year, less than 10% of [0:02] For most of the year, less than 10% of ChatGPT users even visited another one [0:05] ChatGPT users even visited another one [0:05] ChatGPT users even visited another one of the big LLM providers [0:06] of the big LLM providers [0:06] of the big LLM providers >> you open Gemini, it has a pop-up that [0:09] >> you open Gemini, it has a pop-up that [0:09] >> you open Gemini, it has a pop-up that says, "We got Nano Banana. Would you [0:10] says, "We got Nano Banana. Would you [0:10] says, "We got Nano Banana. Would you like to do something with it?" A little [0:12] like to do something with it?" A little [0:12] like to do something with it?" A little pain where you have to type something. [0:13] pain where you have to type something. [0:13] pain where you have to type something. Yeah. I don't I don't know what to do. [0:15] Yeah. I don't I don't know what to do. [0:15] Yeah. I don't I don't know what to do. Yeah. These are product nuances that I I [0:17] Yeah. These are product nuances that I I [0:17] Yeah. These are product nuances that I I think makes people actually take the [0:19] think makes people actually take the [0:19] think makes people actually take the first step. [0:20] first step. [0:20] first step. >> The models have gotten to the level of [0:21] >> The models have gotten to the level of [0:21] >> The models have gotten to the level of quality that you can build a real [0:24] quality that you can build a real [0:24] quality that you can build a real scalable app on top of them. And so the [0:26] scalable app on top of them. And so the [0:26] scalable app on top of them. And so the hope is 2026 will be a huge year for [0:29] hope is 2026 will be a huge year for [0:29] hope is 2026 will be a huge year for consumer builders. [0:36] Today we're talking about who won [0:36] Today we're talking about who won consumer AI in 2025. This was arguably [0:39] consumer AI in 2025. This was arguably [0:39] consumer AI in 2025. This was arguably the year that we saw the big model [0:41] the year that we saw the big model [0:41] the year that we saw the big model providers, OpenAI and Google, most out [0:43] providers, OpenAI and Google, most out [0:44] providers, OpenAI and Google, most out of everyone make a major push of their [0:46] of everyone make a major push of their [0:46] of everyone make a major push of their own into consumer, both in terms of new [0:48] own into consumer, both in terms of new [0:48] own into consumer, both in terms of new models they released, but also in terms [0:50] models they released, but also in terms [0:50] models they released, but also in terms of new products, features, and [0:51] of new products, features, and [0:51] of new products, features, and interfaces that target the mainstream [0:53] interfaces that target the mainstream [0:53] interfaces that target the mainstream user. [0:55] user. [0:55] user. Um you might wonder, why does it matter [0:57] Um you might wonder, why does it matter [0:57] Um you might wonder, why does it matter who is in the lead here? There's some [0:59] who is in the lead here? There's some [0:59] who is in the lead here? There's some early signs that the general LLM [1:01] early signs that the general LLM [1:01] early signs that the general LLM assistant space might be trending [1:03] assistant space might be trending [1:03] assistant space might be trending towards winner take all or at least [1:05] towards winner take all or at least [1:05] towards winner take all or at least winner take most. So only 9% of [1:08] winner take most. So only 9% of [1:08] winner take most. So only 9% of consumers are paying for more than one [1:10] consumers are paying for more than one [1:10] consumers are paying for more than one out of the group of ChatGPT, Gemini, [1:13] out of the group of ChatGPT, Gemini, [1:13] out of the group of ChatGPT, Gemini, Claude, and Cursor. And for most of the [1:15] Claude, and Cursor. And for most of the [1:15] Claude, and Cursor. And for most of the year, less than 10% of ChatGPT users [1:18] year, less than 10% of ChatGPT users [1:18] year, less than 10% of ChatGPT users even visited another one of the big LLM [1:20] even visited another one of the big LLM [1:20] even visited another one of the big LLM providers like Gemini. [1:22] providers like Gemini. [1:22] providers like Gemini. If we had to call it now, ChatGPT is [1:25] If we had to call it now, ChatGPT is [1:25] If we had to call it now, ChatGPT is currently in the lead by far at 800 to [1:28] currently in the lead by far at 800 to [1:28] currently in the lead by far at 800 to 900 million weekly active users. [1:31] 900 million weekly active users. [1:31] 900 million weekly active users. Gemini's added an estimated 35% of their [1:33] Gemini's added an estimated 35% of their [1:33] Gemini's added an estimated 35% of their scale on web and about 40% on mobile. [1:36] scale on web and about 40% on mobile. [1:36] scale on web and about 40% on mobile. And everyone else significantly trails [1:38] And everyone else significantly trails [1:38] And everyone else significantly trails this. So Claude, Grok, Perplexity are [1:40] this. So Claude, Grok, Perplexity are [1:40] this. So Claude, Grok, Perplexity are all about 8 to 10% of the usage. [1:43] all about 8 to 10% of the usage. [1:43] all about 8 to 10% of the usage. But especially in the last 3 to 6 [1:46] But especially in the last 3 to 6 [1:46] But especially in the last 3 to 6 months, things are changing very [1:47] months, things are changing very [1:47] months, things are changing very quickly. With the launch of new viral [1:50] quickly. With the launch of new viral [1:50] quickly. With the launch of new viral models like Nano Banana, Gemini is now [1:52] models like Nano Banana, Gemini is now [1:52] models like Nano Banana, Gemini is now growing desktop users 155% [1:55] growing desktop users 155% [1:55] growing desktop users 155% year-over-year, which is actually [1:56] year-over-year, which is actually [1:56] year-over-year, which is actually accelerating even as they reach more [1:58] accelerating even as they reach more [1:58] accelerating even as they reach more scale, which is pretty crazy to see. And [2:01] scale, which is pretty crazy to see. And [2:01] scale, which is pretty crazy to see. And ChatGPT is only growing 23% [2:03] ChatGPT is only growing 23% [2:03] ChatGPT is only growing 23% year-over-year. [2:04] year-over-year. [2:04] year-over-year. And we're starting to see players like [2:06] And we're starting to see players like [2:06] And we're starting to see players like Anthropic almost specialize within [2:09] Anthropic almost specialize within [2:09] Anthropic almost specialize within consumer owning different verticals like [2:11] consumer owning different verticals like [2:11] consumer owning different verticals like the hyper-technical user. So today we've [2:13] the hyper-technical user. So today we've [2:13] the hyper-technical user. So today we've brought together the A16Z consumer team [2:15] brought together the A16Z consumer team [2:15] brought together the A16Z consumer team to recap what we saw this year from the [2:17] to recap what we saw this year from the [2:17] to recap what we saw this year from the big model companies in consumer and also [2:19] big model companies in consumer and also [2:19] big model companies in consumer and also to predict what might be ahead of us in [2:21] to predict what might be ahead of us in [2:21] to predict what might be ahead of us in 2026. Cool. Well, thank you, Olivia. [2:23] 2026. Cool. Well, thank you, Olivia. [2:23] 2026. Cool. Well, thank you, Olivia. It's It's been a super fun year. If we [2:25] It's It's been a super fun year. If we [2:25] It's It's been a super fun year. If we kind of wind the timeline back to last [2:27] kind of wind the timeline back to last [2:27] kind of wind the timeline back to last January, maybe we should start with what [2:28] January, maybe we should start with what [2:28] January, maybe we should start with what we saw launches, products, what worked, [2:30] we saw launches, products, what worked, [2:30] we saw launches, products, what worked, what didn't. So Justine, tell us what [2:32] what didn't. So Justine, tell us what [2:32] what didn't. So Justine, tell us what you saw this year. OpenAI, Google, what [2:34] you saw this year. OpenAI, Google, what [2:34] you saw this year. OpenAI, Google, what are you paying attention to? What have [2:35] are you paying attention to? What have [2:35] are you paying attention to? What have you changed your mind on? Yeah, those [2:37] you changed your mind on? Yeah, those [2:37] you changed your mind on? Yeah, those two in particular had a ton of consumer [2:38] two in particular had a ton of consumer [2:38] two in particular had a ton of consumer launches, like Olivia mentioned. Um from [2:41] launches, like Olivia mentioned. Um from [2:41] launches, like Olivia mentioned. Um from a model perspective, I would say I would [2:43] a model perspective, I would say I would [2:43] a model perspective, I would say I would argue their most viral models this year, [2:46] argue their most viral models this year, [2:46] argue their most viral models this year, at least among consumers, were in image [2:48] at least among consumers, were in image [2:48] at least among consumers, were in image and video. So for OpenAI, it was the [2:50] and video. So for OpenAI, it was the [2:50] and video. So for OpenAI, it was the ChatGPT 4.0 image, the Ghibli moment, [2:53] ChatGPT 4.0 image, the Ghibli moment, [2:53] ChatGPT 4.0 image, the Ghibli moment, which is crazy that that was this year. [2:55] which is crazy that that was this year. [2:55] which is crazy that that was this year. >> It's crazy that this was this year. It [2:56] >> It's crazy that this was this year. It [2:56] >> It's crazy that this was this year. It feels like it was years ago. [3:02] Um and then Sora, obviously, Sora 2. And [3:02] Um and then Sora, obviously, Sora 2. And then for Google, uh it's V 0, V 0 3, and [3:05] then for Google, uh it's V 0, V 0 3, and [3:05] then for Google, uh it's V 0, V 0 3, and V 0 3.1. Um and then Nano Banana and [3:08] V 0 3.1. Um and then Nano Banana and [3:08] V 0 3.1. Um and then Nano Banana and Nano Banana Pro in image models, which [3:10] Nano Banana Pro in image models, which [3:10] Nano Banana Pro in image models, which went insanely viral, probably comparable [3:13] went insanely viral, probably comparable [3:13] went insanely viral, probably comparable to if not beyond um the the Ghibli [3:15] to if not beyond um the the Ghibli [3:15] to if not beyond um the the Ghibli moment for OpenAI. [3:17] moment for OpenAI. [3:17] moment for OpenAI. I think in terms of the product layer, [3:20] I think in terms of the product layer, [3:20] I think in terms of the product layer, what we saw was OpenAI tended to keep [3:22] what we saw was OpenAI tended to keep [3:22] what we saw was OpenAI tended to keep more things in the ChatGPT interface. So [3:25] more things in the ChatGPT interface. So [3:25] more things in the ChatGPT interface. So like polls, group chats, shopping, [3:29] like polls, group chats, shopping, [3:29] like polls, group chats, shopping, research, tasks, all of these features [3:31] research, tasks, all of these features [3:31] research, tasks, all of these features launched inside ChatGPT as the core. Um [3:34] launched inside ChatGPT as the core. Um [3:34] launched inside ChatGPT as the core. Um the exception there is obviously Sora as [3:36] the exception there is obviously Sora as [3:36] the exception there is obviously Sora as a standalone video app. Whereas Google [3:38] a standalone video app. Whereas Google [3:38] a standalone video app. Whereas Google tended to launch more things as [3:40] tended to launch more things as [3:40] tended to launch more things as standalone products. So they did ship a [3:42] standalone products. So they did ship a [3:42] standalone products. So they did ship a lot through like Google AI Studio um and [3:45] lot through like Google AI Studio um and [3:45] lot through like Google AI Studio um and Google Labs and Gemini and the plethora [3:48] Google Labs and Gemini and the plethora [3:48] Google Labs and Gemini and the plethora of Google services there are to launch a [3:50] of Google services there are to launch a [3:50] of Google services there are to launch a product. Um but they would also ship [3:52] product. Um but they would also ship [3:52] product. Um but they would also ship things as standalone websites that you [3:54] things as standalone websites that you [3:54] things as standalone websites that you could go to and visit, which basically [3:56] could go to and visit, which basically [3:56] could go to and visit, which basically allowed for a more custom interface for [3:58] allowed for a more custom interface for [3:58] allowed for a more custom interface for each type of product, not just the kind [4:00] each type of product, not just the kind [4:00] each type of product, not just the kind of chat entry, chat exit, or image video [4:03] of chat entry, chat exit, or image video [4:03] of chat entry, chat exit, or image video exit. [4:04] exit. [4:04] exit. Well, so Justine, I have a question for [4:05] Well, so Justine, I have a question for [4:05] Well, so Justine, I have a question for you on that. So it felt like 18 months [4:06] you on that. So it felt like 18 months [4:06] you on that. So it felt like 18 months ago we were talking about Midjourney, [4:08] ago we were talking about Midjourney, [4:08] ago we were talking about Midjourney, and most of the multimodal models were [4:10] and most of the multimodal models were [4:10] and most of the multimodal models were defined by aesthetics and realism. Is [4:13] defined by aesthetics and realism. Is [4:13] defined by aesthetics and realism. Is that still true? Like what changed this [4:14] that still true? Like what changed this [4:14] that still true? Like what changed this year? [4:15] year? [4:15] year? Yeah, I think there was definitely [4:17] Yeah, I think there was definitely [4:17] Yeah, I think there was definitely different styles still. And I think [4:18] different styles still. And I think [4:18] different styles still. And I think Midjourney, when you talk to people [4:20] Midjourney, when you talk to people [4:20] Midjourney, when you talk to people really deep in image and video, it still [4:21] really deep in image and video, it still [4:21] really deep in image and video, it still kind of stands apart for this like [4:23] kind of stands apart for this like [4:23] kind of stands apart for this like aesthetic sensibility that a lot of the [4:25] aesthetic sensibility that a lot of the [4:25] aesthetic sensibility that a lot of the models don't have if you don't know how [4:26] models don't have if you don't know how [4:26] models don't have if you don't know how to prompt for it. But I would say this [4:29] to prompt for it. But I would say this [4:29] to prompt for it. But I would say this year in particular we made a lot more [4:30] year in particular we made a lot more [4:30] year in particular we made a lot more strides on realism and also on reasoning [4:34] strides on realism and also on reasoning [4:34] strides on realism and also on reasoning within both image and video. Like all of [4:37] within both image and video. Like all of [4:37] within both image and video. Like all of the little details that make an image or [4:39] the little details that make an image or [4:39] the little details that make an image or a video actually seem real. For example, [4:41] a video actually seem real. For example, [4:41] a video actually seem real. For example, if you have a person walking and [4:42] if you have a person walking and [4:42] if you have a person walking and talking, the people and the cars in the [4:44] talking, the people and the cars in the [4:44] talking, the people and the cars in the background, if they're on a street, [4:46] background, if they're on a street, [4:46] background, if they're on a street, should be moving in the correct [4:47] should be moving in the correct [4:47] should be moving in the correct direction. Like they shouldn't be [4:49] direction. Like they shouldn't be [4:49] direction. Like they shouldn't be morphing and looking strange. Um in [4:51] morphing and looking strange. Um in [4:51] morphing and looking strange. Um in image we were able to, you know, have [4:54] image we were able to, you know, have [4:54] image we were able to, you know, have multiple input images and text and sort [4:56] multiple input images and text and sort [4:56] multiple input images and text and sort of reason across all of those uploads to [4:59] of reason across all of those uploads to [4:59] of reason across all of those uploads to create a cohesive design or something [5:01] create a cohesive design or something [5:01] create a cohesive design or something like that, um which was not something we [5:03] like that, um which was not something we [5:03] like that, um which was not something we saw happening last year for sure. [5:05] saw happening last year for sure. [5:05] saw happening last year for sure. Yeah, I remember when we were excited [5:07] Yeah, I remember when we were excited [5:07] Yeah, I remember when we were excited about having a letter show up correctly [5:10] about having a letter show up correctly [5:10] about having a letter show up correctly in images. And now we have insane [5:13] in images. And now we have insane [5:13] in images. And now we have insane infographics. [5:14] infographics. [5:14] infographics. >> Yes. We can just put up an amazing [5:16] >> Yes. We can just put up an amazing [5:16] >> Yes. We can just put up an amazing YouTube video and say, "Give me an image [5:17] YouTube video and say, "Give me an image [5:17] YouTube video and say, "Give me an image that image that explains this." Yeah. [5:19] that image that explains this." Yeah. [5:19] that image that explains this." Yeah. That's incredibly different. Nano Banana [5:21] That's incredibly different. Nano Banana [5:21] That's incredibly different. Nano Banana Pro can even generate like market maps. [5:23] Pro can even generate like market maps. [5:23] Pro can even generate like market maps. Like I would tell it generate a market [5:25] Like I would tell it generate a market [5:25] Like I would tell it generate a market map. [5:25] map. [5:25] map. >> map. It's incredible. And it will like [5:27] >> map. It's incredible. And it will like [5:27] >> map. It's incredible. And it will like it either has or will go do the web [5:29] it either has or will go do the web [5:29] it either has or will go do the web research within the image model, which [5:30] research within the image model, which [5:30] research within the image model, which is crazy, to get the correct list of [5:32] is crazy, to get the correct list of [5:32] is crazy, to get the correct list of companies and then pull their correct [5:33] companies and then pull their correct [5:33] companies and then pull their correct logos, which is insane. I know. [5:36] logos, which is insane. I know. [5:36] logos, which is insane. I know. There's one benchmark left that the [5:37] There's one benchmark left that the [5:37] There's one benchmark left that the reasoning image models have not cracked. [5:39] reasoning image models have not cracked. [5:39] reasoning image models have not cracked. Not I tested GPT image 1.5 yesterday. [5:42] Not I tested GPT image 1.5 yesterday. [5:42] Not I tested GPT image 1.5 yesterday. They sometimes struggle with like both [5:45] They sometimes struggle with like both [5:45] They sometimes struggle with like both reasoning and multi-step reasoning. So [5:47] reasoning and multi-step reasoning. So [5:47] reasoning and multi-step reasoning. So what I've been testing is you upload a [5:48] what I've been testing is you upload a [5:48] what I've been testing is you upload a picture of an Monopoly board and you [5:50] picture of an Monopoly board and you [5:50] picture of an Monopoly board and you say, "Remove the names of all the [5:52] say, "Remove the names of all the [5:52] say, "Remove the names of all the properties and replace them with names [5:54] properties and replace them with names [5:55] properties and replace them with names of AI labs and startups." And um GPT [5:59] of AI labs and startups." And um GPT [5:59] of AI labs and startups." And um GPT image 1.5 is actually the closest, but [6:02] image 1.5 is actually the closest, but [6:02] image 1.5 is actually the closest, but it's very hard for them to [6:05] it's very hard for them to [6:05] it's very hard for them to do all of those steps. Remove it, come [6:07] do all of those steps. Remove it, come [6:07] do all of those steps. Remove it, come up with the new names, put all of the [6:08] up with the new names, put all of the [6:08] up with the new names, put all of the new names in the correct places, make [6:10] new names in the correct places, make [6:10] new names in the correct places, make sure there's not overlaps or one thing [6:12] sure there's not overlaps or one thing [6:12] sure there's not overlaps or one thing you mentioned three times and another [6:13] you mentioned three times and another [6:14] you mentioned three times and another big player you never mentioned. So [6:15] big player you never mentioned. So [6:15] big player you never mentioned. So there's still some room to go on the [6:17] there's still some room to go on the [6:17] there's still some room to go on the image eval. But it's interesting um that [6:19] image eval. But it's interesting um that [6:19] image eval. But it's interesting um that especially from the image model from [6:21] especially from the image model from [6:21] especially from the image model from ChatGPT, where you can actually see [6:24] ChatGPT, where you can actually see [6:24] ChatGPT, where you can actually see perseverant perseverance of like it it [6:26] perseverant perseverance of like it it [6:26] perseverant perseverance of like it it carries a character over into multiple [6:29] carries a character over into multiple [6:29] carries a character over into multiple image generation, the same style. Yeah. [6:31] image generation, the same style. Yeah. [6:31] image generation, the same style. Yeah. And I thought that was like, "Oh, like [6:33] And I thought that was like, "Oh, like [6:33] And I thought that was like, "Oh, like this is actually very interesting. We're [6:34] this is actually very interesting. We're [6:34] this is actually very interesting. We're storyboarding." Totally. And once it [6:37] storyboarding." Totally. And once it [6:37] storyboarding." Totally. And once it makes you want to generate more. Yeah. [6:39] makes you want to generate more. Yeah. [6:39] makes you want to generate more. Yeah. You know, for me it felt like the most [6:41] You know, for me it felt like the most [6:41] You know, for me it felt like the most underhyped aspect of Nano Banana was the [6:43] underhyped aspect of Nano Banana was the [6:43] underhyped aspect of Nano Banana was the integration with search. Because it [6:45] integration with search. Because it [6:45] integration with search. Because it feels like there's realism, you know, [6:47] feels like there's realism, you know, [6:47] feels like there's realism, you know, which is physics and sort of other [6:49] which is physics and sort of other [6:49] which is physics and sort of other things that feel, you know, we're like [6:50] things that feel, you know, we're like [6:50] things that feel, you know, we're like we're on uncanny valley. There is [6:52] we're on uncanny valley. There is [6:52] we're on uncanny valley. There is reasoning, which is apply modifications [6:55] reasoning, which is apply modifications [6:55] reasoning, which is apply modifications that are adherent to what the user asked [6:57] that are adherent to what the user asked [6:57] that are adherent to what the user asked for. But then there's also a sort of [6:58] for. But then there's also a sort of [6:58] for. But then there's also a sort of accuracy. And for me a good example of [7:01] accuracy. And for me a good example of [7:01] accuracy. And for me a good example of this is product photography. If you say, [7:02] this is product photography. If you say, [7:02] this is product photography. If you say, "Hey, generate a photo of this album [7:04] "Hey, generate a photo of this album [7:04] "Hey, generate a photo of this album cover or a historically accurate photo [7:06] cover or a historically accurate photo [7:06] cover or a historically accurate photo of this moment in time," you have to [7:08] of this moment in time," you have to [7:08] of this moment in time," you have to actually have the search integration, [7:09] actually have the search integration, [7:09] actually have the search integration, and that was sort of non-intuitive, but [7:11] and that was sort of non-intuitive, but [7:11] and that was sort of non-intuitive, but is actually very useful. Totally. Yeah. [7:14] is actually very useful. Totally. Yeah. [7:14] is actually very useful. Totally. Yeah. It's kind of like the V 0 3 moment when [7:16] It's kind of like the V 0 3 moment when [7:16] It's kind of like the V 0 3 moment when I don't think it was intuitive to people [7:18] I don't think it was intuitive to people [7:18] I don't think it was intuitive to people that video would be cracked necessarily [7:19] that video would be cracked necessarily [7:20] that video would be cracked necessarily by bringing audio together with video in [7:21] by bringing audio together with video in [7:21] by bringing audio together with video in the same place, and that ended up being [7:23] the same place, and that ended up being [7:23] the same place, and that ended up being the thing that made AI video go viral. [7:26] the thing that made AI video go viral. [7:26] the thing that made AI video go viral. Like since V 0 3, and and now Sora maybe [7:28] Like since V 0 3, and and now Sora maybe [7:28] Like since V 0 3, and and now Sora maybe dominates, but like since V 0 3, my [7:31] dominates, but like since V 0 3, my [7:31] dominates, but like since V 0 3, my social feeds have been like full of [7:34] social feeds have been like full of [7:34] social feeds have been like full of really realistic-looking AI videos all [7:35] really realistic-looking AI videos all [7:35] really realistic-looking AI videos all the time. [7:36] the time. [7:36] the time. >> 1/5 of my feeds are AI-generated. [7:38] >> 1/5 of my feeds are AI-generated. [7:38] >> 1/5 of my feeds are AI-generated. Amazing. Yeah. What What do you guys do? [7:40] Amazing. Yeah. What What do you guys do? [7:40] Amazing. Yeah. What What do you guys do? There's so many launches this year, and [7:41] There's so many launches this year, and [7:41] There's so many launches this year, and many of them went well, like V 0 and [7:43] many of them went well, like V 0 and [7:43] many of them went well, like V 0 and Nano. Um what do you think is underhyped [7:45] Nano. Um what do you think is underhyped [7:45] Nano. Um what do you think is underhyped or products that you think didn't get [7:46] or products that you think didn't get [7:46] or products that you think didn't get enough attention? [7:48] enough attention? [7:48] enough attention? Brian? [7:49] Brian? [7:49] Brian? It's a good question. [7:53] I think [7:53] I think uh underhyped pulse of the world is [7:56] uh underhyped pulse of the world is [7:56] uh underhyped pulse of the world is probably still underhyped, and you know, [7:58] probably still underhyped, and you know, [7:58] probably still underhyped, and you know, we're talking about [7:59] we're talking about [7:59] we're talking about OpenAI, Google, which to me fall under [8:02] OpenAI, Google, which to me fall under [8:02] OpenAI, Google, which to me fall under productivity category. [8:04] productivity category. [8:04] productivity category. So if you actually think about if you go [8:05] So if you actually think about if you go [8:05] So if you actually think about if you go to App Store today, [8:07] to App Store today, [8:07] to App Store today, top five out of top 10 productivity apps [8:09] top five out of top 10 productivity apps [8:09] top five out of top 10 productivity apps are all Google. [8:10] are all Google. [8:10] are all Google. The it's insane. And and ChatGPT is [8:12] The it's insane. And and ChatGPT is [8:12] The it's insane. And and ChatGPT is number one. So we're talking about a [8:14] number one. So we're talking about a [8:14] number one. So we're talking about a productivity category where it's like [8:15] productivity category where it's like [8:15] productivity category where it's like helps you do things, and I feel like a [8:18] helps you do things, and I feel like a [8:18] helps you do things, and I feel like a lot of people are trying this from [8:20] lot of people are trying this from [8:20] lot of people are trying this from different angle, like how do I actually [8:21] different angle, like how do I actually [8:22] different angle, like how do I actually ingest your data or your schedule, your [8:24] ingest your data or your schedule, your [8:24] ingest your data or your schedule, your email to make it more helpful and give [8:26] email to make it more helpful and give [8:26] email to make it more helpful and give more proactive and notification to you. [8:29] more proactive and notification to you. [8:29] more proactive and notification to you. I think a lot of people are working on [8:30] I think a lot of people are working on [8:30] I think a lot of people are working on it. Given the frequency of people using [8:33] it. Given the frequency of people using [8:33] it. Given the frequency of people using ChatGPT, which I think is what, 25 [8:36] ChatGPT, which I think is what, 25 [8:36] ChatGPT, which I think is what, 25 uh times a week? [8:38] uh times a week? [8:38] uh times a week? Pretty good. Pretty good, three to four [8:40] Pretty good. Pretty good, three to four [8:40] Pretty good. Pretty good, three to four times a day. It feels like it's a really [8:42] times a day. It feels like it's a really [8:42] times a day. It feels like it's a really good position to actually give you [8:44] good position to actually give you [8:44] good position to actually give you proactive nudges and summary and help [8:48] proactive nudges and summary and help [8:48] proactive nudges and summary and help your life in general. So [8:50] your life in general. So [8:50] your life in general. So feel [snorts] like um the everything app [8:52] feel [snorts] like um the everything app [8:52] feel [snorts] like um the everything app was always this myth in in in the [8:54] was always this myth in in in the [8:54] was always this myth in in in the Western world. [8:56] Western world. [8:56] Western world. Think uh OpenAI is trying to move in [8:58] Think uh OpenAI is trying to move in [8:58] Think uh OpenAI is trying to move in that direction where it's ingesting [8:59] that direction where it's ingesting [8:59] that direction where it's ingesting enough people are going there enough [9:02] enough people are going there enough [9:02] enough people are going there enough to start giving really useful proactive [9:04] to start giving really useful proactive [9:04] to start giving really useful proactive nudges, and I think that's a space that [9:06] nudges, and I think that's a space that [9:06] nudges, and I think that's a space that I'm excited about. It's interesting. But [9:08] I'm excited about. It's interesting. But [9:08] I'm excited about. It's interesting. But are you a DAU? I am not a DAU. A pulse? [9:11] are you a DAU? I am not a DAU. A pulse? [9:11] are you a DAU? I am not a DAU. A pulse? Uh not not not a pulse. Similarly, I [9:13] Uh not not not a pulse. Similarly, I [9:13] Uh not not not a pulse. Similarly, I tried pulse for a while and have kind of [9:16] tried pulse for a while and have kind of [9:16] tried pulse for a while and have kind of largely turned off of it. But I would [9:18] largely turned off of it. But I would [9:18] largely turned off of it. But I would agree with you that I feel like pulse [9:20] agree with you that I feel like pulse [9:20] agree with you that I feel like pulse and a couple other examples that OpenAI [9:22] and a couple other examples that OpenAI [9:22] and a couple other examples that OpenAI launched this year are kind of new [9:23] launched this year are kind of new [9:23] launched this year are kind of new primitives or ideas that feel [9:25] primitives or ideas that feel [9:25] primitives or ideas that feel underhyped, but because the execution is [9:28] underhyped, but because the execution is [9:28] underhyped, but because the execution is a little off. [9:29] a little off. [9:29] a little off. >> execution. The usage is off. Another [9:31] >> execution. The usage is off. Another [9:31] >> execution. The usage is off. Another example that I would give, which is [9:32] example that I would give, which is [9:32] example that I would give, which is similarly like personal contacts, would [9:35] similarly like personal contacts, would [9:35] similarly like personal contacts, would be their connectors. So now you can and [9:37] be their connectors. So now you can and [9:37] be their connectors. So now you can and you can do this on Claude as well, you [9:39] you can do this on Claude as well, you [9:39] you can do this on Claude as well, you can connect your calendar, your email, [9:40] can connect your calendar, your email, [9:40] can connect your calendar, your email, your documents. And so hypothetically, [9:42] your documents. And so hypothetically, [9:42] your documents. And so hypothetically, you could say to ChatGPT, you know, read [9:45] you could say to ChatGPT, you know, read [9:45] you could say to ChatGPT, you know, read all of my memos over the past 6 months [9:47] all of my memos over the past 6 months [9:47] all of my memos over the past 6 months and like summarize what's most [9:49] and like summarize what's most [9:49] and like summarize what's most interesting, least interesting. I think [9:50] interesting, least interesting. I think [9:51] interesting, least interesting. I think when that works, it's really exciting. I [9:52] when that works, it's really exciting. I [9:52] when that works, it's really exciting. I have found it to be a little bit [9:54] have found it to be a little bit [9:54] have found it to be a little bit unreliable so far, but I think as the [9:55] unreliable so far, but I think as the [9:55] unreliable so far, but I think as the models get better, they have a real [9:57] models get better, they have a real [9:58] models get better, they have a real chance to kind of own the prosumer [9:59] chance to kind of own the prosumer [9:59] chance to kind of own the prosumer workspace if they get that right. [10:01] workspace if they get that right. [10:01] workspace if they get that right. >> Prosumer is perfect category cuz they we [10:02] >> Prosumer is perfect category cuz they we [10:02] >> Prosumer is perfect category cuz they we we talk about this sometime, but 99% of [10:05] we talk about this sometime, but 99% of [10:05] we talk about this sometime, but 99% of people don't run their life on calendar. [10:07] people don't run their life on calendar. [10:07] people don't run their life on calendar. We do, but [10:09] We do, but [10:09] We do, but but so that that's what I'm thinking [10:10] but so that that's what I'm thinking [10:10] but so that that's what I'm thinking about actual average frequency of using [10:13] about actual average frequency of using [10:14] about actual average frequency of using ChatGPT. And if it's look, if it's 24 [10:16] ChatGPT. And if it's look, if it's 24 [10:16] ChatGPT. And if it's look, if it's 24 times a week, that's pretty good place [10:19] times a week, that's pretty good place [10:19] times a week, that's pretty good place to start. [10:21] to start. [10:21] to start. Olivia, I feel like you are the ultimate [10:22] Olivia, I feel like you are the ultimate [10:22] Olivia, I feel like you are the ultimate power user. You know, what are you still [10:24] power user. You know, what are you still [10:24] power user. You know, what are you still using? What's your stack? It's a great [10:26] using? What's your stack? It's a great [10:26] using? What's your stack? It's a great question. From all of the larger model [10:29] question. From all of the larger model [10:29] question. From all of the larger model companies, actually, I would have to say [10:31] companies, actually, I would have to say [10:31] companies, actually, I would have to say the thing that I'm still using the most [10:32] the thing that I'm still using the most [10:32] the thing that I'm still using the most and was maybe the most impressed by this [10:34] and was maybe the most impressed by this [10:34] and was maybe the most impressed by this year was the Perplexity Comet browser. [10:36] year was the Perplexity Comet browser. [10:37] year was the Perplexity Comet browser. And I don't and was not using Perplexity [10:40] And I don't and was not using Perplexity [10:40] And I don't and was not using Perplexity as my core general LLM assistant. I use [10:42] as my core general LLM assistant. I use [10:42] as my core general LLM assistant. I use ChatGPT and Claude much more, but I [10:46] ChatGPT and Claude much more, but I [10:46] ChatGPT and Claude much more, but I think they really executed on it in a [10:47] think they really executed on it in a [10:47] think they really executed on it in a first-class way in terms of both the [10:49] first-class way in terms of both the [10:49] first-class way in terms of both the agentic model within the browser, but [10:51] agentic model within the browser, but [10:51] agentic model within the browser, but also perhaps more importantly all of the [10:53] also perhaps more importantly all of the [10:53] also perhaps more importantly all of the workflows that you can set up that allow [10:55] workflows that you can set up that allow [10:55] workflows that you can set up that allow you to basically run the same task over [10:58] you to basically run the same task over [10:58] you to basically run the same task over and over either at a preset time or when [11:00] and over either at a preset time or when [11:00] and over either at a preset time or when you trigger it on a certain webpage. So [11:02] you trigger it on a certain webpage. So [11:02] you trigger it on a certain webpage. So that to me was a really exciting launch. [11:03] that to me was a really exciting launch. [11:03] that to me was a really exciting launch. And if you look at the data, like the [11:05] And if you look at the data, like the [11:05] And if you look at the data, like the spike at launch and the sustained [11:07] spike at launch and the sustained [11:07] spike at launch and the sustained traffic for Comet was actually much [11:09] traffic for Comet was actually much [11:09] traffic for Comet was actually much higher than for ChatGPT's own browser [11:12] higher than for ChatGPT's own browser [11:12] higher than for ChatGPT's own browser launch, Atlas, which is kind of crazy [11:14] launch, Atlas, which is kind of crazy [11:14] launch, Atlas, which is kind of crazy given how much more distribution ChatGPT [11:16] given how much more distribution ChatGPT [11:16] given how much more distribution ChatGPT has than Perplexity. Um but I think they [11:19] has than Perplexity. Um but I think they [11:19] has than Perplexity. Um but I think they also launched an email assistant this [11:20] also launched an email assistant this [11:20] also launched an email assistant this year, Perplexity did, and they made a [11:22] year, Perplexity did, and they made a [11:22] year, Perplexity did, and they made a couple acquisitions of really strong [11:24] couple acquisitions of really strong [11:25] couple acquisitions of really strong agentic startups. And so what I would [11:26] agentic startups. And so what I would [11:26] agentic startups. And so what I would love to see from them next year is like [11:28] love to see from them next year is like [11:28] love to see from them next year is like more of these dedicated prosumer [11:30] more of these dedicated prosumer [11:30] more of these dedicated prosumer interfaces. I feel like that would be an [11:32] interfaces. I feel like that would be an [11:32] interfaces. I feel like that would be an awesome direction for them to kind of [11:33] awesome direction for them to kind of [11:33] awesome direction for them to kind of double down in. They do feel like the [11:35] double down in. They do feel like the [11:35] double down in. They do feel like the startup that has the biggest breadth of [11:37] startup that has the biggest breadth of [11:37] startup that has the biggest breadth of ambition, you know, alongside the labs [11:39] ambition, you know, alongside the labs [11:39] ambition, you know, alongside the labs and sort of big tech. Like it's very [11:41] and sort of big tech. Like it's very [11:41] and sort of big tech. Like it's very very impressive just the number of [11:42] very impressive just the number of [11:42] very impressive just the number of things they've shipped this year. Yes, [11:44] things they've shipped this year. Yes, [11:44] things they've shipped this year. Yes, definitely. [11:45] definitely. [11:45] definitely. What I'm you know, one one thing I [11:46] What I'm you know, one one thing I [11:46] What I'm you know, one one thing I wanted to ask you just seen was sort of [11:48] wanted to ask you just seen was sort of [11:48] wanted to ask you just seen was sort of Gemini feels like it's having a real [11:50] Gemini feels like it's having a real [11:50] Gemini feels like it's having a real moment because of all the image and [11:51] moment because of all the image and [11:51] moment because of all the image and video models. Do you think it can [11:53] video models. Do you think it can [11:53] video models. Do you think it can overtake ChatGPT? Is there truly that [11:55] overtake ChatGPT? Is there truly that [11:55] overtake ChatGPT? Is there truly that much demand for these types of models? [11:58] much demand for these types of models? [11:58] much demand for these types of models? I think yes. So what we what I've seen [12:00] I think yes. So what we what I've seen [12:00] I think yes. So what we what I've seen basically is um there is always nearly [12:04] basically is um there is always nearly [12:04] basically is um there is always nearly infinite demand for like the [12:05] infinite demand for like the [12:05] infinite demand for like the best-in-class image or video model [12:08] best-in-class image or video model [12:08] best-in-class image or video model um because then you have a mix of tons [12:11] um because then you have a mix of tons [12:11] um because then you have a mix of tons of different people seeing it and [12:12] of different people seeing it and [12:12] of different people seeing it and wanting to use it. You have like if [12:14] wanting to use it. You have like if [12:14] wanting to use it. You have like if you're using it professionally, if [12:15] you're using it professionally, if [12:15] you're using it professionally, if you're marketing or in entertainment or [12:17] you're marketing or in entertainment or [12:17] you're marketing or in entertainment or storyboarding or whatever, um you always [12:20] storyboarding or whatever, um you always [12:20] storyboarding or whatever, um you always want to be using what's at the forefront [12:21] want to be using what's at the forefront [12:21] want to be using what's at the forefront of the field. And so you're totally fine [12:23] of the field. And so you're totally fine [12:23] of the field. And so you're totally fine to go somewhere other than ChatGPT and [12:25] to go somewhere other than ChatGPT and [12:25] to go somewhere other than ChatGPT and Sora to to get access to VO. Um even if [12:28] Sora to to get access to VO. Um even if [12:28] Sora to to get access to VO. Um even if you're an everyday consumer, so many new [12:30] you're an everyday consumer, so many new [12:30] you're an everyday consumer, so many new viral trends are created around new [12:32] viral trends are created around new [12:32] viral trends are created around new capabilities of the best-in-class image [12:34] capabilities of the best-in-class image [12:34] capabilities of the best-in-class image and video models. And so that ends up [12:37] and video models. And so that ends up [12:37] and video models. And so that ends up driving users into different products [12:39] driving users into different products [12:39] driving users into different products that they may have never tried before. [12:42] that they may have never tried before. [12:42] that they may have never tried before. Like you might be downloading the Gemini [12:44] Like you might be downloading the Gemini [12:44] Like you might be downloading the Gemini app or accidentally ending up on Google [12:46] app or accidentally ending up on Google [12:46] app or accidentally ending up on Google AI Studio, which I know they're trying [12:48] AI Studio, which I know they're trying [12:48] AI Studio, which I know they're trying to make be more for developers, to use [12:50] to make be more for developers, to use [12:50] to make be more for developers, to use Nano Banana Pro, which a lot of users I [12:52] Nano Banana Pro, which a lot of users I [12:53] Nano Banana Pro, which a lot of users I think experienced in the past couple of [12:55] think experienced in the past couple of [12:55] think experienced in the past couple of months. Yeah. The interesting thing [12:57] months. Yeah. The interesting thing [12:57] months. Yeah. The interesting thing about Gemini to me is like [12:58] about Gemini to me is like [12:58] about Gemini to me is like hypothetically they benefit from the [13:00] hypothetically they benefit from the [13:00] hypothetically they benefit from the massive Google distribution advantage. [13:02] massive Google distribution advantage. [13:02] massive Google distribution advantage. Like if you look at Android, [13:04] Like if you look at Android, [13:04] Like if you look at Android, Gemini is at like 50% of ChatGPT's scale [13:07] Gemini is at like 50% of ChatGPT's scale [13:07] Gemini is at like 50% of ChatGPT's scale on mobile, whereas on iOS it's like 17%. [13:10] on mobile, whereas on iOS it's like 17%. [13:10] on mobile, whereas on iOS it's like 17%. So like clearly something is working [13:12] So like clearly something is working [13:12] So like clearly something is working there. They launched a little Gemini [13:13] there. They launched a little Gemini [13:13] there. They launched a little Gemini widget within Chrome recently that [13:15] widget within Chrome recently that [13:15] widget within Chrome recently that encourages you to use it. They're [13:17] encourages you to use it. They're [13:17] encourages you to use it. They're launching it within Google Docs and [13:18] launching it within Google Docs and [13:18] launching it within Google Docs and Gmail and other things. [13:19] Gmail and other things. [13:19] Gmail and other things. >> Yeah. But I think that most the average [13:22] >> Yeah. But I think that most the average [13:22] >> Yeah. But I think that most the average person is still just using one AI [13:24] person is still just using one AI [13:24] person is still just using one AI product. And ChatGPT is like the Kleenex [13:27] product. And ChatGPT is like the Kleenex [13:27] product. And ChatGPT is like the Kleenex of AI. Like it is the brand [laughter] [13:29] of AI. Like it is the brand [laughter] [13:29] of AI. Like it is the brand [laughter] that has become exactly. [13:31] that has become exactly. [13:31] that has become exactly. And so I think that Gemini still has a [13:34] And so I think that Gemini still has a [13:34] And so I think that Gemini still has a pretty big hurdle to overcome just in [13:36] pretty big hurdle to overcome just in [13:36] pretty big hurdle to overcome just in terms of that. [13:38] terms of that. [13:38] terms of that. But if they keep doing what they're [13:39] But if they keep doing what they're [13:39] But if they keep doing what they're doing on these amazing viral consumer [13:42] doing on these amazing viral consumer [13:43] doing on these amazing viral consumer creative tool launches and model [13:44] creative tool launches and model [13:44] creative tool launches and model launches, like [13:46] launches, like [13:46] launches, like they could get there next year. That [13:47] they could get there next year. That [13:47] they could get there next year. That would think you about this. It's really [13:49] would think you about this. It's really [13:49] would think you about this. It's really interesting when you look at [13:52] interesting when you look at [13:52] interesting when you look at Gemini, which which is everywhere. Yeah. [13:54] Gemini, which which is everywhere. Yeah. [13:54] Gemini, which which is everywhere. Yeah. >> Yeah. But yet nowhere to some extent, [13:56] >> Yeah. But yet nowhere to some extent, [13:56] >> Yeah. But yet nowhere to some extent, right? You don't like you know, when you [13:58] right? You don't like you know, when you [13:58] right? You don't like you know, when you look at the actual usage, people still [14:00] look at the actual usage, people still [14:00] look at the actual usage, people still think of the Kleenex [14:01] think of the Kleenex [14:01] think of the Kleenex and they go to [14:02] and they go to [14:02] and they go to ChatGPT. But the interesting thing also [14:04] ChatGPT. But the interesting thing also [14:04] ChatGPT. But the interesting thing also is on the product sensibility. So this [14:07] is on the product sensibility. So this [14:07] is on the product sensibility. So this morning I had like two panes open, [14:10] morning I had like two panes open, [14:10] morning I had like two panes open, OpenAI's image model and and Google's [14:13] OpenAI's image model and and Google's [14:13] OpenAI's image model and and Google's Gemini and basically use a image [14:15] Gemini and basically use a image [14:15] Gemini and basically use a image functionality. When you open Gemini, [14:18] functionality. When you open Gemini, [14:18] functionality. When you open Gemini, it's a blank screen. [14:20] it's a blank screen. [14:20] it's a blank screen. It has a pop-up that says, we got Nano [14:22] It has a pop-up that says, we got Nano [14:22] It has a pop-up that says, we got Nano Banana. Would you like to do something [14:23] Banana. Would you like to do something [14:23] Banana. Would you like to do something with it? And it's the little pane where [14:26] with it? And it's the little pane where [14:26] with it? And it's the little pane where you have to type something. I don't I [14:28] you have to type something. I don't I [14:28] you have to type something. I don't I don't know what to do. [14:30] don't know what to do. [14:30] don't know what to do. >> [laughter] [14:30] >> [laughter] [14:30] >> [laughter] >> ChatGPT, you go in and it has a very [14:33] >> ChatGPT, you go in and it has a very [14:33] >> ChatGPT, you go in and it has a very TikTok-like style of like here's the [14:35] TikTok-like style of like here's the [14:35] TikTok-like style of like here's the trending themes that you might want to [14:37] trending themes that you might want to [14:37] trending themes that you might want to generate. And you click on I want a [14:39] generate. And you click on I want a [14:39] generate. And you click on I want a sketch pen or whatever. And they just [14:40] sketch pen or whatever. And they just [14:40] sketch pen or whatever. And they just like use one of their picture and it [14:42] like use one of their picture and it [14:42] like use one of their picture and it creates something amazing. And then it [14:44] creates something amazing. And then it [14:44] creates something amazing. And then it says, would you like a holiday card? [14:45] says, would you like a holiday card? [14:45] says, would you like a holiday card? Would you like a blah blah blah blah [14:46] Would you like a blah blah blah blah [14:46] Would you like a blah blah blah blah blah. [14:46] blah. [14:46] blah. These are product nuances that I I think [14:49] These are product nuances that I I think [14:49] These are product nuances that I I think makes people actually take the first [14:51] makes people actually take the first [14:51] makes people actually take the first step to generate it. And then once you [14:53] step to generate it. And then once you [14:53] step to generate it. And then once you have it, you have character consistency. [14:55] have it, you have character consistency. [14:55] have it, you have character consistency. So you keep going. So that's interesting [14:58] So you keep going. So that's interesting [14:58] So you keep going. So that's interesting in that I think OpenAI and ChatGPT has [15:00] in that I think OpenAI and ChatGPT has [15:00] in that I think OpenAI and ChatGPT has proven that there's deeper product [15:03] proven that there's deeper product [15:03] proven that there's deeper product sensibility. [15:05] sensibility. [15:05] sensibility. But then I this is a funny thing, maybe [15:07] But then I this is a funny thing, maybe [15:07] But then I this is a funny thing, maybe maybe a little non-kosher thing to say, [15:08] maybe a little non-kosher thing to say, [15:08] maybe a little non-kosher thing to say, but [15:09] but [15:09] but you know, [15:10] you know, [15:10] you know, >> [laughter] [15:10] >> [laughter] [15:10] >> [laughter] >> I I worked at Snap. So when you look at [15:12] >> I I worked at Snap. So when you look at [15:12] >> I I worked at Snap. So when you look at Meta versus Snap, famously Evan Spiegel [15:15] Meta versus Snap, famously Evan Spiegel [15:15] Meta versus Snap, famously Evan Spiegel was chief product officer at Meta. Yeah. [15:17] was chief product officer at Meta. Yeah. [15:17] was chief product officer at Meta. Yeah. Yeah. Um I wonder Yeah. If there is a [15:20] Yeah. Um I wonder Yeah. If there is a [15:20] Yeah. Um I wonder Yeah. If there is a world where the ChatGPT team Yeah. [15:23] world where the ChatGPT team Yeah. [15:23] world where the ChatGPT team Yeah. innovates on the product front again and [15:24] innovates on the product front again and [15:24] innovates on the product front again and again, [15:25] again, [15:25] again, Google with distribution looks at them [15:27] Google with distribution looks at them [15:27] Google with distribution looks at them like, that's cool. Let's just let's just [15:29] like, that's cool. Let's just let's just [15:29] like, that's cool. Let's just let's just integrate it and keep going. And [15:31] integrate it and keep going. And [15:31] integrate it and keep going. And actually play that game. The interesting [15:33] actually play that game. The interesting [15:33] actually play that game. The interesting thing there is that image is paying just [15:35] thing there is that image is paying just [15:35] thing there is that image is paying just launched yesterday when we're filming [15:37] launched yesterday when we're filming [15:37] launched yesterday when we're filming this in ChatGPT. [15:39] this in ChatGPT. [15:39] this in ChatGPT. Brand new. And it took them like they [15:41] Brand new. And it took them like they [15:41] Brand new. And it took them like they had image models for years and it took [15:43] had image models for years and it took [15:43] had image models for years and it took them that long to come up with a [15:45] them that long to come up with a [15:45] them that long to come up with a separate relatively basic interface for [15:47] separate relatively basic interface for [15:47] separate relatively basic interface for generating images. I would almost argue [15:50] generating images. I would almost argue [15:50] generating images. I would almost argue the application layer companies, like [15:52] the application layer companies, like [15:52] the application layer companies, like the the Creas, the Hydras, the Higgs [15:54] the the Creas, the Hydras, the Higgs [15:54] the the Creas, the Hydras, the Higgs Fields of the world popularized that [15:57] Fields of the world popularized that [15:57] Fields of the world popularized that template format and did it first and did [15:59] template format and did it first and did [15:59] template format and did it first and did it better. And they are ChatGPT's [16:02] it better. And they are ChatGPT's [16:02] it better. And they are ChatGPT's product people and then maybe the [16:04] product people and then maybe the [16:04] product people and then maybe the ChatGPT product people. Yeah. Exactly. [16:06] ChatGPT product people. Yeah. Exactly. [16:06] ChatGPT product people. Yeah. Exactly. >> Flawless. Flawless. [16:08] >> Flawless. Flawless. [16:08] >> Flawless. Flawless. Well, maybe going in a slightly [16:09] Well, maybe going in a slightly [16:09] Well, maybe going in a slightly different direction, BK. I'm very [16:10] different direction, BK. I'm very [16:10] different direction, BK. I'm very curious for your take on OpenAI's social [16:12] curious for your take on OpenAI's social [16:13] curious for your take on OpenAI's social features cuz it does feel like that's [16:15] features cuz it does feel like that's [16:15] features cuz it does feel like that's something that you really have to get [16:16] something that you really have to get [16:16] something that you really have to get product execution right on, but also [16:17] product execution right on, but also [16:17] product execution right on, but also network design. You know, there's some [16:19] network design. You know, there's some [16:19] network design. You know, there's some efforts around Sora 2. We should talk [16:20] efforts around Sora 2. We should talk [16:20] efforts around Sora 2. We should talk about that. There's also group chats [16:22] about that. There's also group chats [16:22] about that. There's also group chats within ChatGPT. You're our sort of [16:24] within ChatGPT. You're our sort of [16:24] within ChatGPT. You're our sort of social guy or have been historically. [16:26] social guy or have been historically. [16:26] social guy or have been historically. Bullish, bearish, where's your head at? [16:28] Bullish, bearish, where's your head at? [16:28] Bullish, bearish, where's your head at? Uh bearish for now. Okay. [laughter] [16:30] Uh bearish for now. Okay. [laughter] [16:30] Uh bearish for now. Okay. [laughter] And the reason to me is twofold. [16:33] And the reason to me is twofold. [16:33] And the reason to me is twofold. Uh historically we we look at sort of it [16:36] Uh historically we we look at sort of it [16:36] Uh historically we we look at sort of it it's funny. I I look at products [16:38] it's funny. I I look at products [16:38] it's funny. I I look at products based on what I call inception theory. [16:40] based on what I call inception theory. [16:40] based on what I call inception theory. You go like three to four layers down to [16:41] You go like three to four layers down to [16:42] You go like three to four layers down to figure out what the one-liner is, which [16:43] figure out what the one-liner is, which [16:43] figure out what the one-liner is, which is like I want my dad to love me. [16:45] is like I want my dad to love me. [16:45] is like I want my dad to love me. And so, you know, [laughter] when they [16:47] And so, you know, [laughter] when they [16:47] And so, you know, [laughter] when they think of a product, they're like [16:49] think of a product, they're like [16:49] think of a product, they're like That's that's for me as well as for a [16:50] That's that's for me as well as for a [16:50] That's that's for me as well as for a lot of you. [16:54] And so I I look at some of the, you [16:54] And so I I look at some of the, you know, [16:55] know, [16:55] know, products like like ChatGPT. Ultimately, [16:58] products like like ChatGPT. Ultimately, [16:58] products like like ChatGPT. Ultimately, when you peel the onion five times, I [16:59] when you peel the onion five times, I [16:59] when you peel the onion five times, I think essentially is it helped me be [17:01] think essentially is it helped me be [17:01] think essentially is it helped me be better. [17:02] better. [17:02] better. Like help me get that information. Help [17:04] Like help me get that information. Help [17:04] Like help me get that information. Help me be more productive. Help me be more [17:05] me be more productive. Help me be more [17:05] me be more productive. Help me be more efficient. [17:06] efficient. [17:06] efficient. And then when I think about social [17:08] And then when I think about social [17:08] And then when I think about social features, Meta, Instagram, what have [17:10] features, Meta, Instagram, what have [17:10] features, Meta, Instagram, what have you, or even TikTok, the two layers of [17:13] you, or even TikTok, the two layers of [17:13] you, or even TikTok, the two layers of information or the you know, emotion [17:15] information or the you know, emotion [17:15] information or the you know, emotion that it's trying to address to me is for [17:17] that it's trying to address to me is for [17:17] that it's trying to address to me is for TikTok, entertain me. I want my clown. [17:20] TikTok, entertain me. I want my clown. [17:20] TikTok, entertain me. I want my clown. Entertain me. [17:21] Entertain me. [17:21] Entertain me. And then the other layer is I'm lonely. [17:23] And then the other layer is I'm lonely. [17:23] And then the other layer is I'm lonely. I want to be seen. I want to connect [17:25] I want to be seen. I want to connect [17:25] I want to be seen. I want to connect with people. [17:26] with people. [17:26] with people. And to me these are pretty two different [17:29] And to me these are pretty two different [17:29] And to me these are pretty two different parallels in the product direction. And [17:32] parallels in the product direction. And [17:32] parallels in the product direction. And OpenAI's product is incredible. It's [17:35] OpenAI's product is incredible. It's [17:35] OpenAI's product is incredible. It's magic. It's amazing. But it's ultimately [17:37] magic. It's amazing. But it's ultimately [17:37] magic. It's amazing. But it's ultimately a see me or or help me category, which [17:40] a see me or or help me category, which [17:40] a see me or or help me category, which essentially is why it's the number one [17:42] essentially is why it's the number one [17:42] essentially is why it's the number one in productivity category. [17:45] in productivity category. [17:45] in productivity category. Now we're trying to take this and shove [17:47] Now we're trying to take this and shove [17:47] Now we're trying to take this and shove it in people's life and say, guys, [17:48] it in people's life and say, guys, [17:48] it in people's life and say, guys, connect. Connect better and like [17:50] connect. Connect better and like [17:50] connect. Connect better and like actually feel like you're being seen. [17:53] actually feel like you're being seen. [17:53] actually feel like you're being seen. And even the group chat function, which [17:55] And even the group chat function, which [17:55] And even the group chat function, which I love. It'll be so good to plan a trip [17:57] I love. It'll be so good to plan a trip [17:57] I love. It'll be so good to plan a trip and like actually have that common pain. [18:00] and like actually have that common pain. [18:00] and like actually have that common pain. Um but I think it still stops at [18:03] Um but I think it still stops at [18:03] Um but I think it still stops at probably end count of two to three [18:05] probably end count of two to three [18:05] probably end count of two to three people planning something in a help me [18:07] people planning something in a help me [18:07] people planning something in a help me way versus oh, I feel like I understand [18:11] way versus oh, I feel like I understand [18:11] way versus oh, I feel like I understand Anish so much better because I I've sort [18:13] Anish so much better because I I've sort [18:13] Anish so much better because I I've sort of done that. So largely over time I [18:16] of done that. So largely over time I [18:16] of done that. So largely over time I think that's the reason of that [18:18] think that's the reason of that [18:18] think that's the reason of that division. But that is not to say you can [18:20] division. But that is not to say you can [18:20] division. But that is not to say you can build a separate product that completely [18:23] build a separate product that completely [18:23] build a separate product that completely sort of addresses that. [18:25] sort of addresses that. [18:25] sort of addresses that. I think Sora So we talked about group [18:27] I think Sora So we talked about group [18:27] I think Sora So we talked about group chat. Sora 2 was the other big I think [18:29] chat. Sora 2 was the other big I think [18:29] chat. Sora 2 was the other big I think social push this year from all the [18:31] social push this year from all the [18:31] social push this year from all the consumer AI guys. Which was basically [18:33] consumer AI guys. Which was basically [18:33] consumer AI guys. Which was basically like a TikTok feed, but all AI-generated [18:36] like a TikTok feed, but all AI-generated [18:36] like a TikTok feed, but all AI-generated video and you can make cameos of your [18:37] video and you can make cameos of your [18:37] video and you can make cameos of your friends. The cameos was a very good bet. [18:40] friends. The cameos was a very good bet. [18:40] friends. The cameos was a very good bet. It was a strong bet. Yeah. And I but I [18:42] It was a strong bet. Yeah. And I but I [18:42] It was a strong bet. Yeah. And I but I think what we've seen is like in the [18:44] think what we've seen is like in the [18:44] think what we've seen is like in the retention data and how we're seeing it [18:46] retention data and how we're seeing it [18:46] retention data and how we're seeing it used is it was massively successful as a [18:48] used is it was massively successful as a [18:48] used is it was massively successful as a creator tool. Like now my feed is [18:51] creator tool. Like now my feed is [18:51] creator tool. Like now my feed is probably 2/3 AI slop if not more. And I [18:54] probably 2/3 AI slop if not more. And I [18:54] probably 2/3 AI slop if not more. And I over 50% of it is now Sora, whereas [18:57] over 50% of it is now Sora, whereas [18:57] over 50% of it is now Sora, whereas before it was like all VO and some [18:59] before it was like all VO and some [18:59] before it was like all VO and some claim. Um but it has not been as [19:01] claim. Um but it has not been as [19:01] claim. Um but it has not been as successful as like a social app. [19:04] successful as like a social app. [19:04] successful as like a social app. Yeah. People are like a small number of [19:06] Yeah. People are like a small number of [19:06] Yeah. People are like a small number of creators are creating a ton of content [19:08] creators are creating a ton of content [19:08] creators are creating a ton of content and then bringing it out to like TikTok, [19:10] and then bringing it out to like TikTok, [19:11] and then bringing it out to like TikTok, Instagram, X, Reddit, where it's going [19:13] Instagram, X, Reddit, where it's going [19:13] Instagram, X, Reddit, where it's going massively viral, but it doesn't seem [19:15] massively viral, but it doesn't seem [19:15] massively viral, but it doesn't seem like there's a as much [19:18] like there's a as much [19:18] like there's a as much consumption happening in the app, as [19:20] consumption happening in the app, as [19:20] consumption happening in the app, as much remixing, as much commenting, [19:22] much remixing, as much commenting, [19:22] much remixing, as much commenting, especially as there was initially. You [19:23] especially as there was initially. You [19:23] especially as there was initially. You know, in a funny way the way I think [19:25] know, in a funny way the way I think [19:25] know, in a funny way the way I think about it is like Sora's competition or [19:27] about it is like Sora's competition or [19:27] about it is like Sora's competition or analogy isn't actually TikTok. [19:30] analogy isn't actually TikTok. [19:30] analogy isn't actually TikTok. It's actually CapCut. [19:32] It's actually CapCut. [19:32] It's actually CapCut. Hm. [19:33] Hm. [19:33] Hm. It's a funny way. It's like almost like [19:34] It's a funny way. It's like almost like [19:34] It's a funny way. It's like almost like a creative tool. Yeah. Interesting. Yes. [19:37] a creative tool. Yeah. Interesting. Yes. [19:37] a creative tool. Yeah. Interesting. Yes. I think what I was going to say like I [19:38] I think what I was going to say like I [19:38] I think what I was going to say like I think it goes back to your earlier [19:39] think it goes back to your earlier [19:39] think it goes back to your earlier point, which is like the [19:42] point, which is like the [19:42] point, which is like the kind of motion that drives social apps [19:44] kind of motion that drives social apps [19:44] kind of motion that drives social apps is both these like positive and negative [19:47] is both these like positive and negative [19:47] is both these like positive and negative feelings of like [19:48] feelings of like [19:48] feelings of like oh, I'm publishing this thing of myself [19:51] oh, I'm publishing this thing of myself [19:51] oh, I'm publishing this thing of myself that's kind of sensitive or that I want [19:53] that's kind of sensitive or that I want [19:53] that's kind of sensitive or that I want people to think it's this or that or [19:54] people to think it's this or that or [19:54] people to think it's this or that or this other thing. And so that's kind of [19:56] this other thing. And so that's kind of [19:56] this other thing. And so that's kind of what drives participation on the app. [19:58] what drives participation on the app. [19:58] what drives participation on the app. >> Yeah. The status game. Yeah, it's [20:01] >> Yeah. The status game. Yeah, it's [20:01] >> Yeah. The status game. Yeah, it's exactly the status game. And when it's [20:02] exactly the status game. And when it's [20:02] exactly the status game. And when it's AI generated content and people know [20:04] AI generated content and people know [20:04] AI generated content and people know it's not real, like a real [20:07] it's not real, like a real [20:07] it's not real, like a real representation of you as a human being, [20:09] representation of you as a human being, [20:09] representation of you as a human being, the status game is lost a little bit. [20:11] the status game is lost a little bit. [20:11] the status game is lost a little bit. >> Absolutely lost. Yeah, I think the [20:13] >> Absolutely lost. Yeah, I think the [20:13] >> Absolutely lost. Yeah, I think the status game comes then with [20:15] status game comes then with [20:15] status game comes then with can you prompt something very cool? But [20:17] can you prompt something very cool? But [20:17] can you prompt something very cool? But that's a different type of product and [20:19] that's a different type of product and [20:19] that's a different type of product and that's why I think it goes viral on like [20:21] that's why I think it goes viral on like [20:21] that's why I think it goes viral on like Twitter and all these other existing [20:22] Twitter and all these other existing [20:22] Twitter and all these other existing platforms. I mean my my sort of [20:24] platforms. I mean my my sort of [20:24] platforms. I mean my my sort of counterpoint or bull case for Sora 2 is [20:26] counterpoint or bull case for Sora 2 is [20:26] counterpoint or bull case for Sora 2 is actually think the status game was about [20:28] actually think the status game was about [20:28] actually think the status game was about humor more than anything else. And humor [20:30] humor more than anything else. And humor [20:30] humor more than anything else. And humor is intersection of knowing how to prompt [20:32] is intersection of knowing how to prompt [20:32] is intersection of knowing how to prompt and sort of being culturally aware. [20:34] and sort of being culturally aware. [20:34] and sort of being culturally aware. Yeah. So I think that if they iterated [20:36] Yeah. So I think that if they iterated [20:36] Yeah. So I think that if they iterated on that, that's like a direction that [20:38] on that, that's like a direction that [20:38] on that, that's like a direction that nobody has captured before. Yeah. Yes, [20:40] nobody has captured before. Yeah. Yes, [20:40] nobody has captured before. Yeah. Yes, but if you can export those videos, [20:42] but if you can export those videos, [20:43] but if you can export those videos, isn't it true that like TikTok with Sora [20:46] isn't it true that like TikTok with Sora [20:46] isn't it true that like TikTok with Sora videos on it is strictly better than [20:47] videos on it is strictly better than [20:47] videos on it is strictly better than Sora? We talked about it so much where [20:49] Sora? We talked about it so much where [20:49] Sora? We talked about it so much where like the ultimate [20:51] like the ultimate [20:51] like the ultimate social product is where consumption and [20:53] social product is where consumption and [20:53] social product is where consumption and creation both live together or that the [20:55] creation both live together or that the [20:55] creation both live together or that the output of it is not native to other [20:58] output of it is not native to other [20:58] output of it is not native to other platforms like TikTok, like YouTube [21:01] platforms like TikTok, like YouTube [21:01] platforms like TikTok, like YouTube shorts. [21:02] shorts. [21:02] shorts. So what do folks think of the [21:04] So what do folks think of the [21:04] So what do folks think of the challengers? You know, we're talking [21:05] challengers? You know, we're talking [21:05] challengers? You know, we're talking about Sora 2. I mean Meta, it's crazy to [21:09] about Sora 2. I mean Meta, it's crazy to [21:09] about Sora 2. I mean Meta, it's crazy to talk about Meta as a challenger. I guess [21:10] talk about Meta as a challenger. I guess [21:10] talk about Meta as a challenger. I guess in this context they are, but I think [21:12] in this context they are, but I think [21:12] in this context they are, but I think [laughter] [21:12] [laughter] [21:12] [laughter] Claude, Perplexity, Grok are the more [21:14] Claude, Perplexity, Grok are the more [21:14] Claude, Perplexity, Grok are the more obvious names for challengers. Olivia, [21:16] obvious names for challengers. Olivia, [21:16] obvious names for challengers. Olivia, what's your take? I love Claude. I talk [21:18] what's your take? I love Claude. I talk [21:18] what's your take? I love Claude. I talk to Claude all the time. Claude is so [21:19] to Claude all the time. Claude is so [21:19] to Claude all the time. Claude is so much better place ChatGPT for me as my [21:21] much better place ChatGPT for me as my [21:22] much better place ChatGPT for me as my general LLM. I think Claude is [21:24] general LLM. I think Claude is [21:24] general LLM. I think Claude is opinionated in an interesting way. [21:26] opinionated in an interesting way. [21:26] opinionated in an interesting way. Um I also love Claude because I'm [21:28] Um I also love Claude because I'm [21:28] Um I also love Claude because I'm willing to invest time into building out [21:30] willing to invest time into building out [21:30] willing to invest time into building out AI workflows. I think Claude actually [21:32] AI workflows. I think Claude actually [21:32] AI workflows. I think Claude actually launched a lot of really powerful things [21:34] launched a lot of really powerful things [21:34] launched a lot of really powerful things this year around like artifacts and [21:36] this year around like artifacts and [21:36] this year around like artifacts and skills where you can essentially set up [21:38] skills where you can essentially set up [21:38] skills where you can essentially set up tasks or workflows to run um [21:41] tasks or workflows to run um [21:41] tasks or workflows to run um over time. I do think the reason it [21:43] over time. I do think the reason it [21:43] over time. I do think the reason it hasn't hit the mainstream yet is even [21:45] hasn't hit the mainstream yet is even [21:45] hasn't hit the mainstream yet is even the way they built those things is [21:46] the way they built those things is [21:46] the way they built those things is geared towards a technical user or an [21:49] geared towards a technical user or an [21:49] geared towards a technical user or an engineer. It's I think they tried to [21:52] engineer. It's I think they tried to [21:52] engineer. It's I think they tried to make skills as easy as they could to [21:54] make skills as easy as they could to [21:54] make skills as easy as they could to create and it still was not anywhere [21:57] create and it still was not anywhere [21:57] create and it still was not anywhere near easy enough for the mainstream [21:58] near easy enough for the mainstream [21:58] near easy enough for the mainstream consumer. Um another example would be [22:01] consumer. Um another example would be [22:01] consumer. Um another example would be they were actually the first of the big [22:03] they were actually the first of the big [22:03] they were actually the first of the big players to kind of launch file creation, [22:06] players to kind of launch file creation, [22:06] players to kind of launch file creation, slide deck creation, editing and they [22:09] slide deck creation, editing and they [22:09] slide deck creation, editing and they branded it as like file generation and [22:11] branded it as like file generation and [22:11] branded it as like file generation and analysis or something and it was like a [22:13] analysis or something and it was like a [22:13] analysis or something and it was like a toggle feature within a [22:15] toggle feature within a [22:15] toggle feature within a >> [laughter] [22:15] >> [laughter] [22:15] >> [laughter] >> a setting bar of a setting bar or [22:17] >> a setting bar of a setting bar or [22:17] >> a setting bar of a setting bar or something. So like very few people used [22:19] something. So like very few people used [22:19] something. So like very few people used it. And yet to me it's still the best [22:21] it. And yet to me it's still the best [22:21] it. And yet to me it's still the best product across all of them at doing that [22:23] product across all of them at doing that [22:23] product across all of them at doing that kind of complex work. So I love Claude, [22:26] kind of complex work. So I love Claude, [22:26] kind of complex work. So I love Claude, but I think if they want to be a true [22:29] but I think if they want to be a true [22:29] but I think if they want to be a true mainstream consumer product, they need [22:32] mainstream consumer product, they need [22:32] mainstream consumer product, they need to [22:33] to [22:33] to dumb it down even more in terms of [22:36] dumb it down even more in terms of [22:36] dumb it down even more in terms of accessibility. There was that survey you [22:37] accessibility. There was that survey you [22:37] accessibility. There was that survey you found recently of US teens. Yeah, [22:39] found recently of US teens. Yeah, [22:39] found recently of US teens. Yeah, there's I think it was three times more [22:42] there's I think it was three times more [22:42] there's I think it was three times more US teens have ever used Character AI [22:44] US teens have ever used Character AI [22:44] US teens have ever used Character AI than have used Claude. [22:45] than have used Claude. [22:45] than have used Claude. >> Yeah. So I think that shows that like [22:47] >> Yeah. So I think that shows that like [22:47] >> Yeah. So I think that shows that like Claude is a pretty broad thing. Yeah. [22:49] Claude is a pretty broad thing. Yeah. [22:49] Claude is a pretty broad thing. Yeah. Claude is beloved amongst tech people, [22:51] Claude is beloved amongst tech people, [22:51] Claude is beloved amongst tech people, but outside of tech people, I think they [22:53] but outside of tech people, I think they [22:53] but outside of tech people, I think they are maybe struggling to pick up [22:55] are maybe struggling to pick up [22:55] are maybe struggling to pick up relevance. It is interesting though like [22:57] relevance. It is interesting though like [22:57] relevance. It is interesting though like if you look at the sort of aesthetics, [22:59] if you look at the sort of aesthetics, [22:59] if you look at the sort of aesthetics, the product design, the craft, like [23:01] the product design, the craft, like [23:01] the product design, the craft, like three things that Anthropic did were [23:03] three things that Anthropic did were [23:03] three things that Anthropic did were MCP, skills, and command line interface, [23:07] MCP, skills, and command line interface, [23:07] MCP, skills, and command line interface, Claude code. Like those are three [23:09] Claude code. Like those are three [23:09] Claude code. Like those are three surprising bets, especially Claude code. [23:11] surprising bets, especially Claude code. [23:11] surprising bets, especially Claude code. I would have said command line interface [23:13] I would have said command line interface [23:13] I would have said command line interface really? Like is this the way that people [23:15] really? Like is this the way that people [23:15] really? Like is this the way that people want to interact [23:15] want to interact [23:15] want to interact >> you were going to talk about taking over [23:17] >> you were going to talk about taking over [23:17] >> you were going to talk about taking over air mail and the thinking cap. [23:19] air mail and the thinking cap. [23:19] air mail and the thinking cap. Yeah, that too. [23:21] Yeah, that too. [23:21] Yeah, that too. >> [laughter] [23:24] >> So three things like where's the [23:24] >> So three things like where's the thinking cap? [23:25] thinking cap? [23:25] thinking cap? But it's sort of very high-minded [23:27] But it's sort of very high-minded [23:27] But it's sort of very high-minded design. I don't know if it's sort of [23:28] design. I don't know if it's sort of [23:29] design. I don't know if it's sort of like versus market or or maybe that's [23:31] like versus market or or maybe that's [23:31] like versus market or or maybe that's apologetic on their behalf, but I think [23:33] apologetic on their behalf, but I think [23:33] apologetic on their behalf, but I think it is it is that it's opinionated and [23:35] it is it is that it's opinionated and [23:35] it is it is that it's opinionated and it's great. [23:35] it's great. [23:35] it's great. >> Yeah, yeah. I do need to hear Justine's [23:37] >> Yeah, yeah. I do need to hear Justine's [23:37] >> Yeah, yeah. I do need to hear Justine's take on both Meta and Grok as I feel [23:39] take on both Meta and Grok as I feel [23:39] take on both Meta and Grok as I feel like they both had fascinating years. [23:42] like they both had fascinating years. [23:42] like they both had fascinating years. Yes. So Meta [23:45] Yes. So Meta [23:45] Yes. So Meta hired all those researchers. I think [23:46] hired all those researchers. I think [23:46] hired all those researchers. I think their strongest models are actually not [23:48] their strongest models are actually not [23:48] their strongest models are actually not consumer facing models. It's their SAM 3 [23:51] consumer facing models. It's their SAM 3 [23:51] consumer facing models. It's their SAM 3 series. So like the segment anything for [23:54] series. So like the segment anything for [23:54] series. So like the segment anything for video, for image, and for audio. And [23:56] video, for image, and for audio. And [23:56] video, for image, and for audio. And basically like for video for example, [23:59] basically like for video for example, [23:59] basically like for video for example, you can upload a video and you can [24:00] you can upload a video and you can [24:00] you can upload a video and you can describe a natural language like find [24:03] describe a natural language like find [24:03] describe a natural language like find the kid in the red t-shirt and it will [24:05] the kid in the red t-shirt and it will [24:05] the kid in the red t-shirt and it will find and track that person across every [24:08] find and track that person across every [24:08] find and track that person across every the entire video even if they're coming [24:10] the entire video even if they're coming [24:10] the entire video even if they're coming in and out of the frame. It will let you [24:12] in and out of the frame. It will let you [24:12] in and out of the frame. It will let you apply effects like blurring them out or [24:14] apply effects like blurring them out or [24:14] apply effects like blurring them out or removing them or whatever. And you can [24:16] removing them or whatever. And you can [24:16] removing them or whatever. And you can imagine a similar thing with audio [24:18] imagine a similar thing with audio [24:18] imagine a similar thing with audio with different stems and then with image [24:21] with different stems and then with image [24:21] with different stems and then with image with different objects in an image. I [24:23] with different objects in an image. I [24:23] with different objects in an image. I think we're going to see next year [24:25] think we're going to see next year [24:25] think we're going to see next year hopefully some incredible consumer [24:26] hopefully some incredible consumer [24:26] hopefully some incredible consumer products built on top of those models, [24:29] products built on top of those models, [24:29] products built on top of those models, but today they're more of a playground [24:31] but today they're more of a playground [24:31] but today they're more of a playground for developers than they are a consumer [24:33] for developers than they are a consumer [24:33] for developers than they are a consumer facing product. [24:34] facing product. [24:34] facing product. >> given just like the DNA of the company. [24:37] >> given just like the DNA of the company. [24:37] >> given just like the DNA of the company. Yeah, there so the one good consumer [24:39] Yeah, there so the one good consumer [24:39] Yeah, there so the one good consumer feature I think they've launched this [24:40] feature I think they've launched this [24:40] feature I think they've launched this year and with AI is the Instagram AI [24:42] year and with AI is the Instagram AI [24:43] year and with AI is the Instagram AI translations [24:44] translations [24:44] translations where [clears throat] [24:45] where [clears throat] [24:45] where [clears throat] when you're uploading a reel now, you [24:47] when you're uploading a reel now, you [24:47] when you're uploading a reel now, you can opt in to enable translations and it [24:49] can opt in to enable translations and it [24:49] can opt in to enable translations and it will clone your voice, translate it into [24:52] will clone your voice, translate it into [24:52] will clone your voice, translate it into five different languages, [24:54] five different languages, [24:54] five different languages, apply the translation with your voice. [24:56] apply the translation with your voice. [24:56] apply the translation with your voice. So you know, and then re-dub with with [24:58] So you know, and then re-dub with with [24:58] So you know, and then re-dub with with the lip sync. [24:59] the lip sync. [24:59] the lip sync. >> Wow. And so it basically makes it seem [25:01] >> Wow. And so it basically makes it seem [25:01] >> Wow. And so it basically makes it seem like you're a native speaker in in [25:02] like you're a native speaker in in [25:02] like you're a native speaker in in whatever language. So I would love to [25:04] whatever language. So I would love to [25:04] whatever language. So I would love to see more of that stuff come to to the [25:06] see more of that stuff come to to the [25:06] see more of that stuff come to to the Meta products. [25:09] Meta products. [25:09] Meta products. Grok I think has had So Grok had a crazy [25:11] Grok I think has had So Grok had a crazy [25:11] Grok I think has had So Grok had a crazy year with like the companions, with all [25:14] year with like the companions, with all [25:14] year with like the companions, with all of the LLM progress and the coding [25:15] of the LLM progress and the coding [25:15] of the LLM progress and the coding progress. [25:17] progress. [25:17] progress. I think their image and video progress [25:19] I think their image and video progress [25:19] I think their image and video progress is probably the steepest slope I've seen [25:22] is probably the steepest slope I've seen [25:22] is probably the steepest slope I've seen of any of the companies. Like [25:25] of any of the companies. Like [25:25] of any of the companies. Like it was probably like six months ago they [25:27] it was probably like six months ago they [25:27] it was probably like six months ago they didn't even like have image and video [25:28] didn't even like have image and video [25:28] didn't even like have image and video models and they're shipping so fast to [25:32] models and they're shipping so fast to [25:32] models and they're shipping so fast to launch new features. Like it was [25:33] launch new features. Like it was [25:33] launch new features. Like it was initially just image to video, they [25:35] initially just image to video, they [25:35] initially just image to video, they added text to video, they added audio, [25:37] added text to video, they added audio, [25:37] added text to video, they added audio, then they added lip sync with speech, [25:39] then they added lip sync with speech, [25:39] then they added lip sync with speech, then they added 15 second videos. Like [25:42] then they added 15 second videos. Like [25:42] then they added 15 second videos. Like they're just not slowing down the speed [25:44] they're just not slowing down the speed [25:44] they're just not slowing down the speed of progress and Elon has made a bunch of [25:46] of progress and Elon has made a bunch of [25:46] of progress and Elon has made a bunch of statements about like wanting more [25:48] statements about like wanting more [25:48] statements about like wanting more interactive video game type content out [25:50] interactive video game type content out [25:50] interactive video game type content out of Grok and wanting movies out of Grok [25:52] of Grok and wanting movies out of Grok [25:52] of Grok and wanting movies out of Grok by the end of next year. So [25:55] by the end of next year. So [25:55] by the end of next year. So let's hope it continues to go at that [25:56] let's hope it continues to go at that [25:56] let's hope it continues to go at that pace. Do you feel like it's a pincer [25:58] pace. Do you feel like it's a pincer [25:58] pace. Do you feel like it's a pincer movement where like on one hand there's [26:00] movement where like on one hand there's [26:00] movement where like on one hand there's like a very [26:01] like a very [26:01] like a very infrastructural model layer of like [26:03] infrastructural model layer of like [26:03] infrastructural model layer of like let's get to the let's top the LLM arena [26:06] let's get to the let's top the LLM arena [26:06] let's get to the let's top the LLM arena charts. And then the other one is like [26:09] charts. And then the other one is like [26:09] charts. And then the other one is like let's go Annie. [26:11] let's go Annie. [26:11] let's go Annie. I think That's like a little bit of like [26:13] I think That's like a little bit of like [26:13] I think That's like a little bit of like a bifurcated move. Right, like the [26:15] a bifurcated move. Right, like the [26:15] a bifurcated move. Right, like the entertainment and the like smart [26:16] entertainment and the like smart [26:17] entertainment and the like smart >> Absolutely, but entertainment in a way [26:18] >> Absolutely, but entertainment in a way [26:18] >> Absolutely, but entertainment in a way that like we're talking about, you know, [26:20] that like we're talking about, you know, [26:20] that like we're talking about, you know, Anthropic and ChatGPT's general [26:22] Anthropic and ChatGPT's general [26:22] Anthropic and ChatGPT's general population, but you just said Character [26:24] population, but you just said Character [26:24] population, but you just said Character AI is way more popular. [26:26] AI is way more popular. [26:26] AI is way more popular. >> Yes. So then like how do we think about [26:28] >> Yes. So then like how do we think about [26:28] >> Yes. So then like how do we think about that? And I think, you know, it's it's a [26:29] that? And I think, you know, it's it's a [26:29] that? And I think, you know, it's it's a very interesting strategy in my mind. [26:32] very interesting strategy in my mind. [26:32] very interesting strategy in my mind. And Grok like in the image and video [26:34] And Grok like in the image and video [26:34] And Grok like in the image and video app, since pretty early on they've had [26:36] app, since pretty early on they've had [26:36] app, since pretty early on they've had templates of popular things like you're [26:39] templates of popular things like you're [26:39] templates of popular things like you're standing somewhere and suddenly like a [26:41] standing somewhere and suddenly like a [26:41] standing somewhere and suddenly like a thing drops a rope drops from the [26:42] thing drops a rope drops from the [26:42] thing drops a rope drops from the ceiling and you grab onto it and it like [26:44] ceiling and you grab onto it and it like [26:44] ceiling and you grab onto it and it like swings you out of the scene. Like some [26:46] swings you out of the scene. Like some [26:46] swings you out of the scene. Like some really good ones that go viral regularly [26:48] really good ones that go viral regularly [26:48] really good ones that go viral regularly on TikTok and other places. Yeah. [26:51] on TikTok and other places. Yeah. [26:51] on TikTok and other places. Yeah. Really, really interesting. Well, so [26:52] Really, really interesting. Well, so [26:52] Really, really interesting. Well, so maybe switching gears from 25 to 26, [26:56] maybe switching gears from 25 to 26, [26:56] maybe switching gears from 25 to 26, what are some of all your predictions [26:57] what are some of all your predictions [26:57] what are some of all your predictions for next year? What do you think we'll [26:58] for next year? What do you think we'll [26:58] for next year? What do you think we'll see hardware, models, commerce we [27:01] see hardware, models, commerce we [27:02] see hardware, models, commerce we haven't spoken about yet? So what do we [27:04] haven't spoken about yet? So what do we [27:04] haven't spoken about yet? So what do we think we'll play out? [27:06] think we'll play out? [27:06] think we'll play out? I think I know this is we're talking [27:07] I think I know this is we're talking [27:07] I think I know this is we're talking about consumer, but one of the things [27:09] about consumer, but one of the things [27:09] about consumer, but one of the things that's been really maybe underrated for [27:11] that's been really maybe underrated for [27:11] that's been really maybe underrated for me about ChatGPT that we might see more [27:13] me about ChatGPT that we might see more [27:13] me about ChatGPT that we might see more of next year is they've really made a [27:15] of next year is they've really made a [27:15] of next year is they've really made a push into the enterprise both with the [27:17] push into the enterprise both with the [27:18] push into the enterprise both with the traditional enterprise licenses and then [27:19] traditional enterprise licenses and then [27:19] traditional enterprise licenses and then working with specific companies even [27:21] working with specific companies even [27:21] working with specific companies even like train models for them. And I think [27:23] like train models for them. And I think [27:23] like train models for them. And I think when we think about the fact that most [27:26] when we think about the fact that most [27:26] when we think about the fact that most consumers only use one general LLM [27:28] consumers only use one general LLM [27:28] consumers only use one general LLM product, [27:29] product, [27:29] product, ChatGPT enterprise usage, they published [27:31] ChatGPT enterprise usage, they published [27:32] ChatGPT enterprise usage, they published a big study, but it's up somewhat like [27:33] a big study, but it's up somewhat like [27:33] a big study, but it's up somewhat like eight or nine X year over year. Yeah. [27:35] eight or nine X year over year. Yeah. [27:35] eight or nine X year over year. Yeah. And so for entering a world now where [27:37] And so for entering a world now where [27:37] And so for entering a world now where people have to use ChatGPT for their [27:41] people have to use ChatGPT for their [27:41] people have to use ChatGPT for their company or as part of their work, that [27:43] company or as part of their work, that [27:43] company or as part of their work, that could really translate into consumer [27:44] could really translate into consumer [27:45] could really translate into consumer usage [27:46] usage [27:46] usage or maybe they become the the workspace [27:49] or maybe they become the the workspace [27:49] or maybe they become the the workspace with the connectors and some of the [27:50] with the connectors and some of the [27:50] with the connectors and some of the other things that they're investing in [27:52] other things that they're investing in [27:52] other things that they're investing in and someone [snorts] else owns the [27:54] and someone [snorts] else owns the [27:54] and someone [snorts] else owns the consumer consumer use cases. I think to [27:57] consumer consumer use cases. I think to [27:57] consumer consumer use cases. I think to that end we have to talk about their [27:59] that end we have to talk about their [27:59] that end we have to talk about their push into apps and I think whether or [28:01] push into apps and I think whether or [28:01] push into apps and I think whether or not that works is going to be kind of [28:02] not that works is going to be kind of [28:02] not that works is going to be kind of the defining question for them next [28:04] the defining question for them next [28:04] the defining question for them next year. Yeah, and I think that the we've [28:06] year. Yeah, and I think that the we've [28:06] year. Yeah, and I think that the we've all discussed the importance of the apps [28:08] all discussed the importance of the apps [28:08] all discussed the importance of the apps SDK and the apps directory as they're [28:10] SDK and the apps directory as they're [28:10] SDK and the apps directory as they're calling it and it's going to be a huge [28:11] calling it and it's going to be a huge [28:11] calling it and it's going to be a huge new channel for consumer. I think what's [28:13] new channel for consumer. I think what's [28:13] new channel for consumer. I think what's less discussed is it's hyper relevant to [28:14] less discussed is it's hyper relevant to [28:14] less discussed is it's hyper relevant to enterprise. So I think where ChatGPT [28:17] enterprise. So I think where ChatGPT [28:17] enterprise. So I think where ChatGPT shines is where it's able to operate [28:19] shines is where it's able to operate [28:19] shines is where it's able to operate across a number of tools for one [28:21] across a number of tools for one [28:21] across a number of tools for one workflow. And if you think about the [28:22] workflow. And if you think about the [28:22] workflow. And if you think about the number of things you do in your sort of [28:24] number of things you do in your sort of [28:24] number of things you do in your sort of business day-to-day that operates across [28:26] business day-to-day that operates across [28:26] business day-to-day that operates across many tools, it's most of those things. [28:28] many tools, it's most of those things. [28:28] many tools, it's most of those things. So I think that will have very [28:29] So I think that will have very [28:29] So I think that will have very interesting implications for the SAS [28:31] interesting implications for the SAS [28:31] interesting implications for the SAS ecosystem and it's a part of the app [28:33] ecosystem and it's a part of the app [28:33] ecosystem and it's a part of the app store we're not talking about as much. [28:35] store we're not talking about as much. [28:35] store we're not talking about as much. Yeah. Yeah, [28:36] Yeah. Yeah, [28:36] Yeah. Yeah, maybe less of a prediction, but thinking [28:38] maybe less of a prediction, but thinking [28:38] maybe less of a prediction, but thinking through 2025 and we talked about all the [28:41] through 2025 and we talked about all the [28:41] through 2025 and we talked about all the big move from big labs and [28:44] big move from big labs and [28:44] big move from big labs and from the startup point, I think one of [28:46] from the startup point, I think one of [28:46] from the startup point, I think one of the biggest trend we've seen is app [28:48] the biggest trend we've seen is app [28:48] the biggest trend we've seen is app generation. [28:49] generation. [28:49] generation. Mhm. And and I think there's a real [28:52] Mhm. And and I think there's a real [28:52] Mhm. And and I think there's a real world where [28:53] world where [28:53] world where we see the big labs with a distribution [28:56] we see the big labs with a distribution [28:56] we see the big labs with a distribution and the frequency of usage of people [28:58] and the frequency of usage of people [28:58] and the frequency of usage of people coming in to start saying, look like [29:00] coming in to start saying, look like [29:00] coming in to start saying, look like maybe there is a common [29:02] maybe there is a common [29:02] maybe there is a common type of product and apps that we could [29:04] type of product and apps that we could [29:05] type of product and apps that we could actually help you generate within the [29:06] actually help you generate within the [29:06] actually help you generate within the confines of [29:08] confines of [29:08] confines of the big lab products. [29:09] the big lab products. [29:09] the big lab products. >> Yeah. I think that's like one of the [29:10] >> Yeah. I think that's like one of the [29:10] >> Yeah. I think that's like one of the interesting thing which, you know, again [29:12] interesting thing which, you know, again [29:12] interesting thing which, you know, again going back to the supply chain of ideas [29:13] going back to the supply chain of ideas [29:13] going back to the supply chain of ideas and research, maybe that's one thing and [29:16] and research, maybe that's one thing and [29:16] and research, maybe that's one thing and again nothing groundbreaking, but as we [29:18] again nothing groundbreaking, but as we [29:18] again nothing groundbreaking, but as we know the Ghibli broke the internet. [29:21] know the Ghibli broke the internet. [29:22] know the Ghibli broke the internet. My cousin who who knows nothing about [29:25] My cousin who who knows nothing about [29:25] My cousin who who knows nothing about tech sent me a Ghibli photo. [29:28] tech sent me a Ghibli photo. [29:28] tech sent me a Ghibli photo. Um Well, let's not send this to your [29:29] Um Well, let's not send this to your [29:29] Um Well, let's not send this to your cousin then. [29:32] cousin then. [29:32] cousin then. >> [laughter] [29:33] >> [laughter] [29:33] >> [laughter] >> And I think that goes to show that [29:35] >> And I think that goes to show that [29:35] >> And I think that goes to show that templates matter, that style matters. [29:38] templates matter, that style matters. [29:39] templates matter, that style matters. >> Yeah. [29:39] >> Yeah. [29:39] >> Yeah. And I think about video and like it's [29:41] And I think about video and like it's [29:41] And I think about video and like it's pretty freaking good. [29:43] pretty freaking good. [29:43] pretty freaking good. >> Yeah. [29:43] >> Yeah. [29:43] >> Yeah. And it's possible that we're already at [29:46] And it's possible that we're already at [29:46] And it's possible that we're already at a point that it's not necessarily just [29:48] a point that it's not necessarily just [29:48] a point that it's not necessarily just about the capability of models of the [29:50] about the capability of models of the [29:50] about the capability of models of the big labs, but the stylistic things, the [29:53] big labs, but the stylistic things, the [29:53] big labs, but the stylistic things, the template. Think of TikTok. [29:55] template. Think of TikTok. [29:55] template. Think of TikTok. The large capability largely still the [29:57] The large capability largely still the [29:58] The large capability largely still the same. Music, trend, dance, go. [30:01] same. Music, trend, dance, go. [30:01] same. Music, trend, dance, go. Except the trend and format keeps on [30:03] Except the trend and format keeps on [30:03] Except the trend and format keeps on changing and keeps it extremely fresh. [30:05] changing and keeps it extremely fresh. [30:05] changing and keeps it extremely fresh. So, I feel like there's a real world [30:07] So, I feel like there's a real world [30:07] So, I feel like there's a real world where [30:08] where [30:08] where the repurpose or a team or what have you [30:09] the repurpose or a team or what have you [30:09] the repurpose or a team or what have you can start thinking about ways to [30:11] can start thinking about ways to [30:11] can start thinking about ways to actually really build in [30:13] actually really build in [30:13] actually really build in video first products, entities, live [30:15] video first products, entities, live [30:16] video first products, entities, live models, and I think the cost will go [30:17] models, and I think the cost will go [30:17] models, and I think the cost will go down enough for people to try it out, [30:19] down enough for people to try it out, [30:19] down enough for people to try it out, and I'm excited to see that. [30:21] and I'm excited to see that. [30:21] and I'm excited to see that. Yeah, I think what I'm most excited [30:22] Yeah, I think what I'm most excited [30:22] Yeah, I think what I'm most excited about is sort of along those lines, [30:25] about is sort of along those lines, [30:25] about is sort of along those lines, basically everything becoming [30:26] basically everything becoming [30:26] basically everything becoming multimodal. Like I call it like anything [30:29] multimodal. Like I call it like anything [30:29] multimodal. Like I call it like anything in to anything out, which is basically [30:32] in to anything out, which is basically [30:32] in to anything out, which is basically initially especially with these image [30:33] initially especially with these image [30:33] initially especially with these image and video models, [30:35] and video models, [30:35] and video models, it was you put in a text prompt and you [30:38] it was you put in a text prompt and you [30:38] it was you put in a text prompt and you get an image out or a video out. You [30:41] get an image out or a video out. You [30:41] get an image out or a video out. You couldn't really do much with it. And now [30:44] couldn't really do much with it. And now [30:44] couldn't really do much with it. And now we've started to see this with the image [30:46] we've started to see this with the image [30:46] we've started to see this with the image edit models with like Nano Banana and [30:48] edit models with like Nano Banana and [30:48] edit models with like Nano Banana and with Flux and and with the new OpenAI [30:51] with Flux and and with the new OpenAI [30:51] with Flux and and with the new OpenAI model where you can put an image in now [30:54] model where you can put an image in now [30:54] model where you can put an image in now and get an another image out. You can [30:57] and get an another image out. You can [30:57] and get an another image out. You can put an image in with a text pair in a [30:58] put an image in with a text pair in a [30:58] put an image in with a text pair in a direction or put an image with a [31:00] direction or put an image with a [31:00] direction or put an image with a template and another reference image and [31:02] template and another reference image and [31:02] template and another reference image and get another image out. What happens when [31:05] get another image out. What happens when [31:05] get another image out. What happens when you can put a video in and get images [31:07] you can put a video in and get images [31:07] you can put a video in and get images out that are related to or the the next [31:10] out that are related to or the the next [31:10] out that are related to or the the next iteration of the video. Or you can put a [31:12] iteration of the video. Or you can put a [31:12] iteration of the video. Or you can put a video in and a text prompt about what [31:15] video in and a text prompt about what [31:15] video in and a text prompt about what you want to edit and get the edited [31:17] you want to edit and get the edited [31:17] you want to edit and get the edited video out. From my conversations with [31:20] video out. From my conversations with [31:20] video out. From my conversations with the labs, a lot of them are trying to [31:22] the labs, a lot of them are trying to [31:22] the labs, a lot of them are trying to basically combine all of these largely [31:24] basically combine all of these largely [31:24] basically combine all of these largely separate efforts they've had across like [31:27] separate efforts they've had across like [31:27] separate efforts they've had across like text reasoning and intelligence, the LLM [31:29] text reasoning and intelligence, the LLM [31:29] text reasoning and intelligence, the LLM space, and image and video into like [31:32] space, and image and video into like [31:32] space, and image and video into like what if we can put merge those all into [31:35] what if we can put merge those all into [31:35] what if we can put merge those all into like a mega model that can take a lot [31:37] like a mega model that can take a lot [31:37] like a mega model that can take a lot different forms of content and and [31:39] different forms of content and and [31:39] different forms of content and and produce much more. I think it's also [31:41] produce much more. I think it's also [31:41] produce much more. I think it's also going to have huge implications for like [31:43] going to have huge implications for like [31:43] going to have huge implications for like design. Cuz if you think about it, a lot [31:46] design. Cuz if you think about it, a lot [31:46] design. Cuz if you think about it, a lot of design is combining images with text [31:49] of design is combining images with text [31:49] of design is combining images with text with video with different elements in [31:51] with video with different elements in [31:51] with video with different elements in kind of interesting ways. Yeah. [31:54] kind of interesting ways. Yeah. [31:54] kind of interesting ways. Yeah. I guess if I think about like a macro [31:55] I guess if I think about like a macro [31:55] I guess if I think about like a macro level prediction, [31:57] level prediction, [31:57] level prediction, I think it's actually going to be more [31:59] I think it's actually going to be more [31:59] I think it's actually going to be more of the same in that when we talk about [32:01] of the same in that when we talk about [32:01] of the same in that when we talk about what all of the [32:02] what all of the [32:02] what all of the labs have launched in consumer, [32:04] labs have launched in consumer, [32:05] labs have launched in consumer, they've done a great job with models and [32:06] they've done a great job with models and [32:06] they've done a great job with models and they've done a great job with [32:08] they've done a great job with [32:08] they've done a great job with incremental things that improve the core [32:09] incremental things that improve the core [32:10] incremental things that improve the core experience of using like a ChatGPT or [32:12] experience of using like a ChatGPT or [32:12] experience of using like a ChatGPT or Gemini. [32:13] Gemini. [32:13] Gemini. In my opinion, we've gone through dozens [32:15] In my opinion, we've gone through dozens [32:15] In my opinion, we've gone through dozens of things that they've launched or tried [32:17] of things that they've launched or tried [32:17] of things that they've launched or tried as new consumer products or new consumer [32:19] as new consumer products or new consumer [32:19] as new consumer products or new consumer interfaces like group chat, like pulse, [32:21] interfaces like group chat, like pulse, [32:22] interfaces like group chat, like pulse, like Atlas, like Sora. Google has had a [32:25] like Atlas, like Sora. Google has had a [32:25] like Atlas, like Sora. Google has had a long tail like stitch, gems, opal, tons. [32:29] long tail like stitch, gems, opal, tons. [32:29] long tail like stitch, gems, opal, tons. >> Yeah. None of those are really working, [32:31] >> Yeah. None of those are really working, [32:31] >> Yeah. None of those are really working, and I think it's because it's not the [32:32] and I think it's because it's not the [32:32] and I think it's because it's not the core competency of these companies [32:35] core competency of these companies [32:35] core competency of these companies anymore to build opinionated standalone [32:37] anymore to build opinionated standalone [32:38] anymore to build opinionated standalone consumer UI. Out of all of those, I [32:40] consumer UI. Out of all of those, I [32:40] consumer UI. Out of all of those, I think the product that's working the [32:42] think the product that's working the [32:42] think the product that's working the most is like Notebook LM, and that's one [32:44] most is like Notebook LM, and that's one [32:44] most is like Notebook LM, and that's one of like maybe 20 things that Google has [32:46] of like maybe 20 things that Google has [32:46] of like maybe 20 things that Google has tried or experimented with. So, I think [32:48] tried or experimented with. So, I think [32:48] tried or experimented with. So, I think it's actually very positive for startups [32:50] it's actually very positive for startups [32:50] it's actually very positive for startups in that consumer startups in that the [32:52] in that consumer startups in that the [32:52] in that consumer startups in that the models will keep getting better, which [32:53] models will keep getting better, which [32:53] models will keep getting better, which the startups can use, [32:55] the startups can use, [32:55] the startups can use, and they'll keep, you know, they'll make [32:56] and they'll keep, you know, they'll make [32:57] and they'll keep, you know, they'll make ChatGPT better and better, but I don't [32:58] ChatGPT better and better, but I don't [32:58] ChatGPT better and better, but I don't necessarily think that ChatGPT like [33:01] necessarily think that ChatGPT like [33:01] necessarily think that ChatGPT like verticalizes into all of these other [33:03] verticalizes into all of these other [33:03] verticalizes into all of these other amazing use cases or products, and [33:05] amazing use cases or products, and [33:05] amazing use cases or products, and there's still room for startups to be [33:06] there's still room for startups to be [33:06] there's still room for startups to be building there. [33:07] building there. [33:07] building there. >> I have a yes and to that where [33:09] >> I have a yes and to that where [33:09] >> I have a yes and to that where absolutely [33:10] absolutely [33:10] absolutely but however, when the input and the [33:12] but however, when the input and the [33:13] but however, when the input and the output is text [33:15] output is text [33:15] output is text Yeah. where ChatGPT and Gemini of the [33:17] Yeah. where ChatGPT and Gemini of the [33:17] Yeah. where ChatGPT and Gemini of the world shine the most, no matter how [33:19] world shine the most, no matter how [33:19] world shine the most, no matter how deeper you go, no matter how specific [33:21] deeper you go, no matter how specific [33:21] deeper you go, no matter how specific you think your text output is going to [33:23] you think your text output is going to [33:23] you think your text output is going to be, essentially given the frequency of [33:25] be, essentially given the frequency of [33:25] be, essentially given the frequency of use users of the main big lab products, [33:29] use users of the main big lab products, [33:29] use users of the main big lab products, I think it's going to be really hard to [33:30] I think it's going to be really hard to [33:30] I think it's going to be really hard to stitch that and get that away from that [33:32] stitch that and get that away from that [33:32] stitch that and get that away from that usage if your product is mainly text in [33:36] usage if your product is mainly text in [33:36] usage if your product is mainly text in text out. Yeah. [33:37] text out. Yeah. [33:37] text out. Yeah. So, I I do think you you have to be [33:39] So, I I do think you you have to be [33:39] So, I I do think you you have to be creative around what is the angle that [33:42] creative around what is the angle that [33:42] creative around what is the angle that you can like go steal people away from. [33:44] you can like go steal people away from. [33:45] you can like go steal people away from. You know, I love that you used the word [33:46] You know, I love that you used the word [33:46] You know, I love that you used the word opinionated because I think that for [33:48] opinionated because I think that for [33:48] opinionated because I think that for labs, certainly for big tech and perhaps [33:50] labs, certainly for big tech and perhaps [33:50] labs, certainly for big tech and perhaps increasingly for labs, the priorities [33:52] increasingly for labs, the priorities [33:52] increasingly for labs, the priorities get set in their promo committee always. [33:54] get set in their promo committee always. [33:54] get set in their promo committee always. And if you're a PM, and it's always the [33:56] And if you're a PM, and it's always the [33:56] And if you're a PM, and it's always the sort of mid-career PMs, and I've been [33:57] sort of mid-career PMs, and I've been [33:57] sort of mid-career PMs, and I've been one of these, and you like the [33:58] one of these, and you like the [33:58] one of these, and you like the incentives are always to get promoted, [34:00] incentives are always to get promoted, [34:00] incentives are always to get promoted, and the way to get promoted is to build [34:02] and the way to get promoted is to build [34:02] and the way to get promoted is to build something safe that extends a core [34:04] something safe that extends a core [34:04] something safe that extends a core metric and a core feature. [34:06] metric and a core feature. [34:06] metric and a core feature. So, building opinionated products is a [34:08] So, building opinionated products is a [34:08] So, building opinionated products is a very risky way to manage your career, [34:10] very risky way to manage your career, [34:10] very risky way to manage your career, you know, cuz they're probably not going [34:11] you know, cuz they're probably not going [34:11] you know, cuz they're probably not going to work. They're probably going to have [34:12] to work. They're probably going to have [34:12] to work. They're probably going to have a bunch of implications for legal and [34:14] a bunch of implications for legal and [34:14] a bunch of implications for legal and compliance, and the CEO might yell at [34:16] compliance, and the CEO might yell at [34:16] compliance, and the CEO might yell at you. So, I just think that they are so [34:18] you. So, I just think that they are so [34:18] you. So, I just think that they are so structured to do incremental things. The [34:20] structured to do incremental things. The [34:20] structured to do incremental things. The more founders do opinionated things, the [34:22] more founders do opinionated things, the [34:22] more founders do opinionated things, the more advantage they are. I think [34:23] more advantage they are. I think [34:24] more advantage they are. I think honestly the big thing we haven't [34:25] honestly the big thing we haven't [34:25] honestly the big thing we haven't discussed here, too, is compute, which [34:26] discussed here, too, is compute, which [34:26] discussed here, too, is compute, which is the labs have this inherent tension [34:28] is the labs have this inherent tension [34:29] is the labs have this inherent tension between [34:30] between [34:30] between there's a limited amount of compute, and [34:31] there's a limited amount of compute, and [34:31] there's a limited amount of compute, and they either spend it on like training [34:33] they either spend it on like training [34:33] they either spend it on like training models or they spend it on inference. [34:35] models or they spend it on inference. [34:35] models or they spend it on inference. And even with inference, there's this [34:37] And even with inference, there's this [34:37] And even with inference, there's this split between like the entertainment [34:39] split between like the entertainment [34:39] split between like the entertainment Ghibli use cases and the like coding [34:41] Ghibli use cases and the like coding [34:41] Ghibli use cases and the like coding intelligence use cases. I think xAI is [34:43] intelligence use cases. I think xAI is [34:43] intelligence use cases. I think xAI is probably the only model company that is [34:45] probably the only model company that is [34:45] probably the only model company that is not bottlenecked on compute from my [34:47] not bottlenecked on compute from my [34:47] not bottlenecked on compute from my understanding, whereas the others have [34:49] understanding, whereas the others have [34:49] understanding, whereas the others have to make really like serious and [34:51] to make really like serious and [34:51] to make really like serious and significant calls of like if we let if [34:53] significant calls of like if we let if [34:53] significant calls of like if we let if we release Nano Banana and go super [34:55] we release Nano Banana and go super [34:55] we release Nano Banana and go super viral, like it may slow down the next [34:57] viral, like it may slow down the next [34:57] viral, like it may slow down the next like big LLM we're trying to push [34:59] like big LLM we're trying to push [34:59] like big LLM we're trying to push forward, whereas startups who focus on [35:01] forward, whereas startups who focus on [35:01] forward, whereas startups who focus on the app layer don't have that problem [35:03] the app layer don't have that problem [35:03] the app layer don't have that problem because there's there's no tension [35:04] because there's there's no tension [35:04] because there's there's no tension there. [35:05] there. [35:05] there. >> Absolutely. Yeah. [35:07] >> Absolutely. Yeah. [35:07] >> Absolutely. Yeah. We've talked about this before. I also [35:08] We've talked about this before. I also [35:08] We've talked about this before. I also think that there are categories in which [35:09] think that there are categories in which [35:09] think that there are categories in which being multimodal is just allows you to [35:12] being multimodal is just allows you to [35:12] being multimodal is just allows you to deliver a better proposition to the [35:13] deliver a better proposition to the [35:14] deliver a better proposition to the customer, and the labs and big tech are [35:16] customer, and the labs and big tech are [35:16] customer, and the labs and big tech are always going to be sort of [35:16] always going to be sort of [35:17] always going to be sort of definitionally first-party model only. [35:19] definitionally first-party model only. [35:19] definitionally first-party model only. So, I think as all the models get [35:20] So, I think as all the models get [35:20] So, I think as all the models get better, perhaps 80% of what you need can [35:22] better, perhaps 80% of what you need can [35:23] better, perhaps 80% of what you need can be received from a single model, but for [35:25] be received from a single model, but for [35:25] be received from a single model, but for the power users, and so much of AI is a [35:27] the power users, and so much of AI is a [35:27] the power users, and so much of AI is a power user story. You know, you always [35:29] power user story. You know, you always [35:29] power user story. You know, you always said that like well, power users are [35:30] said that like well, power users are [35:30] said that like well, power users are just power users, and I think that's [35:32] just power users, and I think that's [35:32] just power users, and I think that's true in a pre-AI world, but now the kind [35:34] true in a pre-AI world, but now the kind [35:34] true in a pre-AI world, but now the kind of depth of value and the depth of [35:35] of depth of value and the depth of [35:36] of depth of value and the depth of monetization is so much higher that [35:37] monetization is so much higher that [35:37] monetization is so much higher that maybe all of AI is actually a power user [35:39] maybe all of AI is actually a power user [35:39] maybe all of AI is actually a power user story. You know, and everyone else is [35:41] story. You know, and everyone else is [35:41] story. You know, and everyone else is just traffic. Yes. [35:42] just traffic. Yes. [35:43] just traffic. Yes. >> Yeah. [35:43] >> Yeah. [35:43] >> Yeah. Which is why we're also seeing like [35:45] Which is why we're also seeing like [35:45] Which is why we're also seeing like consumer products for the first time [35:46] consumer products for the first time [35:46] consumer products for the first time ever have more than 100% revenue [35:48] ever have more than 100% revenue [35:48] ever have more than 100% revenue retention. [35:49] retention. [35:49] retention. >> Yes. And that's separating the good from [35:51] >> Yes. And that's separating the good from [35:51] >> Yes. And that's separating the good from the great from the exceptional in [35:53] the great from the exceptional in [35:53] the great from the exceptional in consumer AI. And to be clear, how that [35:55] consumer AI. And to be clear, how that [35:55] consumer AI. And to be clear, how that happens is [35:56] happens is [35:56] happens is they charge for usage often in addition [35:58] they charge for usage often in addition [35:58] they charge for usage often in addition to a subscription. So, you can use [36:00] to a subscription. So, you can use [36:00] to a subscription. So, you can use beyond whatever your quota is for the [36:02] beyond whatever your quota is for the [36:02] beyond whatever your quota is for the month given your subscription and pay [36:03] month given your subscription and pay [36:03] month given your subscription and pay more. [36:04] more. [36:04] more. >> upgrade of the tier or actually buying [36:06] >> upgrade of the tier or actually buying [36:06] >> upgrade of the tier or actually buying tokens [36:07] tokens [36:07] tokens or or more more usage. Yeah. It's that's [36:10] or or more more usage. Yeah. It's that's [36:10] or or more more usage. Yeah. It's that's what differentiates it. Like, you know, [36:11] what differentiates it. Like, you know, [36:12] what differentiates it. Like, you know, if you told me pre-AI [36:14] if you told me pre-AI [36:14] if you told me pre-AI we see a consumer company with 100 plus [36:16] we see a consumer company with 100 plus [36:16] we see a consumer company with 100 plus retention in money, I'm like that that [36:18] retention in money, I'm like that that [36:18] retention in money, I'm like that that doesn't make any sense. That's not [36:19] doesn't make any sense. That's not [36:19] doesn't make any sense. That's not compute. Yeah. Yeah. [36:21] compute. Yeah. Yeah. [36:21] compute. Yeah. Yeah. No pun intended. Exactly. [36:23] No pun intended. Exactly. [36:23] No pun intended. Exactly. >> [laughter] [36:24] >> [laughter] [36:24] >> [laughter] >> Exactly. Well, guys, okay, maybe let's [36:26] >> Exactly. Well, guys, okay, maybe let's [36:26] >> Exactly. Well, guys, okay, maybe let's talk about start with specific [36:27] talk about start with specific [36:27] talk about start with specific recommendations. Like after this pod, [36:29] recommendations. Like after this pod, [36:29] recommendations. Like after this pod, what are the products people should [36:30] what are the products people should [36:30] what are the products people should download or the features or the models? [36:31] download or the features or the models? [36:31] download or the features or the models? What should folks be using today? [36:37] I I guess on the multimodal point, I [36:37] I I guess on the multimodal point, I think one really [36:39] think one really [36:39] think one really underhyped product that people should [36:40] underhyped product that people should [36:40] underhyped product that people should check out, not because they'll use it [36:42] check out, not because they'll use it [36:42] check out, not because they'll use it every day, but because it shows sort of [36:43] every day, but because it shows sort of [36:44] every day, but because it shows sort of what is possible when you combine like [36:46] what is possible when you combine like [36:46] what is possible when you combine like an agent with image with text is Pameli. [36:49] an agent with image with text is Pameli. [36:49] an agent with image with text is Pameli. So, this is like the Google Labs product [36:51] So, this is like the Google Labs product [36:51] So, this is like the Google Labs product where you put in the URL of your [36:52] where you put in the URL of your [36:52] where you put in the URL of your business, and it has an agent go to the [36:55] business, and it has an agent go to the [36:55] business, and it has an agent go to the website, pull all of the product and [36:57] website, pull all of the product and [36:57] website, pull all of the product and brand photos, summarize what it thinks [36:59] brand photos, summarize what it thinks [36:59] brand photos, summarize what it thinks your brand's aesthetic is, what it [37:02] your brand's aesthetic is, what it [37:02] your brand's aesthetic is, what it stands for, what kind of customers it's [37:04] stands for, what kind of customers it's [37:04] stands for, what kind of customers it's targeting, and then it will generate [37:06] targeting, and then it will generate [37:06] targeting, and then it will generate three different ad campaigns for you. [37:09] three different ad campaigns for you. [37:09] three different ad campaigns for you. And it will generate not only the text, [37:10] And it will generate not only the text, [37:11] And it will generate not only the text, but it will generate like the Instagram [37:12] but it will generate like the Instagram [37:12] but it will generate like the Instagram posts. It will generate the flyer. It [37:14] posts. It will generate the flyer. It [37:14] posts. It will generate the flyer. It will generate like the the photo of your [37:16] will generate like the the photo of your [37:16] will generate like the the photo of your product in this, you know, whatever [37:19] product in this, you know, whatever [37:19] product in this, you know, whatever wherever it thinks it should be based on [37:20] wherever it thinks it should be based on [37:20] wherever it thinks it should be based on your customer. [37:22] your customer. [37:22] your customer. And very cool product. Would be hard to [37:26] And very cool product. Would be hard to [37:26] And very cool product. Would be hard to become a giant standalone product within [37:27] become a giant standalone product within [37:27] become a giant standalone product within Google, I think, but shows sort of the [37:30] Google, I think, but shows sort of the [37:30] Google, I think, but shows sort of the future of what happens if we combine [37:32] future of what happens if we combine [37:32] future of what happens if we combine agents with generation models that have [37:35] agents with generation models that have [37:35] agents with generation models that have sort of really deep understanding of [37:37] sort of really deep understanding of [37:37] sort of really deep understanding of context that an image model or video [37:39] context that an image model or video [37:40] context that an image model or video model normally wouldn't have. [37:41] model normally wouldn't have. [37:41] model normally wouldn't have. Startup products, though. Do you have a [37:43] Startup products, though. Do you have a [37:43] Startup products, though. Do you have a favorite startup product? [37:46] favorite startup product? [37:46] favorite startup product? In creative tools? I think I I mean, [37:48] In creative tools? I think I I mean, [37:48] In creative tools? I think I I mean, we're investors in Krea, so this is [37:49] we're investors in Krea, so this is [37:49] we're investors in Krea, so this is biased, but I think they they have [37:51] biased, but I think they they have [37:52] biased, but I think they they have they've really done an exceptional job [37:53] they've really done an exceptional job [37:53] they've really done an exceptional job of being the best place to use [37:57] of being the best place to use [37:57] of being the best place to use every model or every quality model [37:59] every model or every quality model [37:59] every model or every quality model across every modality, and also building [38:01] across every modality, and also building [38:01] across every modality, and also building more of the interface on top of these [38:03] more of the interface on top of these [38:03] more of the interface on top of these models. Like I now prefer to use Nano [38:05] models. Like I now prefer to use Nano [38:06] models. Like I now prefer to use Nano Banana Pro on Krea because Krea allows [38:08] Banana Pro on Krea because Krea allows [38:08] Banana Pro on Krea because Krea allows you to [clears throat] save elements, [38:10] you to [clears throat] save elements, [38:10] you to [clears throat] save elements, which are essentially characters or [38:12] which are essentially characters or [38:12] which are essentially characters or styles or objects that you can like at [38:14] styles or objects that you can like at [38:14] styles or objects that you can like at tag to reprompt versus having to drag [38:17] tag to reprompt versus having to drag [38:17] tag to reprompt versus having to drag the same image reference into Nano [38:19] the same image reference into Nano [38:19] the same image reference into Nano Banana over and over again. It's a good [38:21] Banana over and over again. It's a good [38:21] Banana over and over again. It's a good one. I suppose it falls under startup [38:24] one. I suppose it falls under startup [38:24] one. I suppose it falls under startup category. Again, shilling companies, but [38:27] category. Again, shilling companies, but [38:27] category. Again, shilling companies, but you know, [38:28] you know, [38:28] you know, the one that I use the most is actually [38:30] the one that I use the most is actually [38:30] the one that I use the most is actually 11 Labs Reader. [38:31] 11 Labs Reader. [38:31] 11 Labs Reader. And the reason is we've seen an [38:33] And the reason is we've seen an [38:33] And the reason is we've seen an explosion on podcasts, and there's I [38:35] explosion on podcasts, and there's I [38:35] explosion on podcasts, and there's I think a reason for that, right? People [38:37] think a reason for that, right? People [38:37] think a reason for that, right? People are a lot more on the go. [38:39] are a lot more on the go. [38:39] are a lot more on the go. The reading [38:40] The reading [38:40] The reading capability of us reading, I think it's [38:42] capability of us reading, I think it's [38:42] capability of us reading, I think it's going down over time. [38:44] going down over time. [38:44] going down over time. And so, you know, let's not fight the [38:45] And so, you know, let's not fight the [38:45] And so, you know, let's not fight the reality. Let's embrace it, and okay, so [38:48] reality. Let's embrace it, and okay, so [38:48] reality. Let's embrace it, and okay, so then like let's actually find a written [38:50] then like let's actually find a written [38:50] then like let's actually find a written material, translate into listening, and [38:53] material, translate into listening, and [38:53] material, translate into listening, and and do that. And I used to be a power [38:55] and do that. And I used to be a power [38:56] and do that. And I used to be a power user of tools like Pocket. You know, I [38:58] user of tools like Pocket. You know, I [38:58] user of tools like Pocket. You know, I didn't have time to read everything that [39:00] didn't have time to read everything that [39:00] didn't have time to read everything that I wanted to read, and it's a saving [39:02] I wanted to read, and it's a saving [39:02] I wanted to read, and it's a saving behavior, right? You're you're going [39:03] behavior, right? You're you're going [39:03] behavior, right? You're you're going around and saving all the things you [39:04] around and saving all the things you [39:04] around and saving all the things you eventually want to consume. [39:06] eventually want to consume. [39:06] eventually want to consume. But I think what I do now is similar [39:08] But I think what I do now is similar [39:08] But I think what I do now is similar where I go get all the things I want to [39:10] where I go get all the things I want to [39:10] where I go get all the things I want to read, and I just put it either PDF it or [39:13] read, and I just put it either PDF it or [39:13] read, and I just put it either PDF it or put it in 11 Reader. And just like once [39:15] put it in 11 Reader. And just like once [39:15] put it in 11 Reader. And just like once in a while when I'm on a walk and I have [39:17] in a while when I'm on a walk and I have [39:17] in a while when I'm on a walk and I have like 3 4 minutes, you know, 1.5x speed [39:19] like 3 4 minutes, you know, 1.5x speed [39:20] like 3 4 minutes, you know, 1.5x speed or 2x speed, and just listen to one of [39:21] or 2x speed, and just listen to one of [39:21] or 2x speed, and just listen to one of these and get the gist of it. So, I [39:23] these and get the gist of it. So, I [39:23] these and get the gist of it. So, I think that's been a good way to [39:25] think that's been a good way to [39:25] think that's been a good way to use a little bit of time as a sort of [39:28] use a little bit of time as a sort of [39:28] use a little bit of time as a sort of semi-normal person. [39:31] semi-normal person. [39:31] semi-normal person. Well, first of all, I love this question [39:32] Well, first of all, I love this question [39:33] Well, first of all, I love this question cuz I am strongly opinionated that by [39:35] cuz I am strongly opinionated that by [39:35] cuz I am strongly opinionated that by far the best way to get up to speed on [39:37] far the best way to get up to speed on [39:37] far the best way to get up to speed on AI is just to try a ton of products, and [39:39] AI is just to try a ton of products, and [39:39] AI is just to try a ton of products, and you get opinionated really quickly. Um, [39:42] you get opinionated really quickly. Um, [39:42] you get opinionated really quickly. Um, and I actually for the whole month of [39:44] and I actually for the whole month of [39:44] and I actually for the whole month of December are on Twitter publishing one [39:46] December are on Twitter publishing one [39:46] December are on Twitter publishing one new consumer product a day for people to [39:48] new consumer product a day for people to [39:48] new consumer product a day for people to check out. So, that's one way. Um, I'll [39:51] check out. So, that's one way. Um, I'll [39:51] check out. So, that's one way. Um, I'll name three others that I think are super [39:54] name three others that I think are super [39:54] name three others that I think are super maybe relevant or interesting that [39:55] maybe relevant or interesting that [39:55] maybe relevant or interesting that people can plug into their workflows. [39:57] people can plug into their workflows. [39:57] people can plug into their workflows. So, one would be Gamma for slide deck [39:59] So, one would be Gamma for slide deck [39:59] So, one would be Gamma for slide deck generations. You can go text prompt to [40:01] generations. You can go text prompt to [40:01] generations. You can go text prompt to slide deck, you can go document to slide [40:03] slide deck, you can go document to slide [40:03] slide deck, you can go document to slide deck. I use it for everything. Also, the [40:06] deck. I use it for everything. Also, the [40:06] deck. I use it for everything. Also, the slides are flexible sizes, so you're no [40:08] slides are flexible sizes, so you're no [40:08] slides are flexible sizes, so you're no longer like editing every little pixel [40:10] longer like editing every little pixel [40:10] longer like editing every little pixel in your Google Slides to get it to fit [40:11] in your Google Slides to get it to fit [40:11] in your Google Slides to get it to fit into one, which is great. [40:13] into one, which is great. [40:13] into one, which is great. Um, Granola for note-taking. You might [40:15] Um, Granola for note-taking. You might [40:15] Um, Granola for note-taking. You might not have any meetings over the holidays, [40:16] not have any meetings over the holidays, [40:17] not have any meetings over the holidays, but in the new year, um, and it just [40:19] but in the new year, um, and it just [40:19] but in the new year, um, and it just gets better and better the more meetings [40:20] gets better and better the more meetings [40:20] gets better and better the more meetings you have on it cuz it has the context of [40:22] you have on it cuz it has the context of [40:22] you have on it cuz it has the context of what you talked about before. And then [40:24] what you talked about before. And then [40:24] what you talked about before. And then lastly, I'm still going to plug try the [40:26] lastly, I'm still going to plug try the [40:26] lastly, I'm still going to plug try the Comet browser. If you want to try kind [40:28] Comet browser. If you want to try kind [40:28] Comet browser. If you want to try kind of an AI-native workspace, I think [40:29] of an AI-native workspace, I think [40:29] of an AI-native workspace, I think that's one of the most accessible ones [40:31] that's one of the most accessible ones [40:31] that's one of the most accessible ones to start with. [40:33] to start with. [40:33] to start with. I mean, for me, I've spent my whole year [40:35] I mean, for me, I've spent my whole year [40:35] I mean, for me, I've spent my whole year obsessed with coding and AI code. Um, [40:37] obsessed with coding and AI code. Um, [40:37] obsessed with coding and AI code. Um, it's just been so tremendously fun. I, [40:39] it's just been so tremendously fun. I, [40:39] it's just been so tremendously fun. I, by the way, Brian would take the other [40:40] by the way, Brian would take the other [40:40] by the way, Brian would take the other side of your argument that the big labs [40:43] side of your argument that the big labs [40:43] side of your argument that the big labs or big tech will win in app generation. [40:45] or big tech will win in app generation. [40:45] or big tech will win in app generation. I think they just lack the focus. [40:47] I think they just lack the focus. [40:47] I think they just lack the focus. Products like Opal have been, you know, [40:49] Products like Opal have been, you know, [40:49] Products like Opal have been, you know, released with a whimper and they're one [40:51] released with a whimper and they're one [40:51] released with a whimper and they're one model only. So, I I didn't say they [40:53] model only. So, I I didn't say they [40:53] model only. So, I I didn't say they would win it. I think that we will see [40:55] would win it. I think that we will see [40:55] would win it. I think that we will see them doing it. Yes, yes. I think that's [40:57] them doing it. Yes, yes. I think that's [40:57] them doing it. Yes, yes. I think that's true, but I think for the pure consumer [40:59] true, but I think for the pure consumer [40:59] true, but I think for the pure consumer side, of course, Wabi is is really fun [41:01] side, of course, Wabi is is really fun [41:01] side, of course, Wabi is is really fun and really capable and I think they are [41:03] and really capable and I think they are [41:03] and really capable and I think they are they're creating the right sort of, um, [41:06] they're creating the right sort of, um, [41:06] they're creating the right sort of, um, constraints on app generation so that [41:08] constraints on app generation so that [41:08] constraints on app generation so that you can get a really satisfying [41:10] you can get a really satisfying [41:10] you can get a really satisfying functional result. And I think so far [41:12] functional result. And I think so far [41:12] functional result. And I think so far there's been a lot of over-promising in [41:13] there's been a lot of over-promising in [41:13] there's been a lot of over-promising in app generation, which has discouraged [41:15] app generation, which has discouraged [41:15] app generation, which has discouraged the early users. I also think if you [41:17] the early users. I also think if you [41:17] the early users. I also think if you haven't tried, you know, GPT-52 in Codex [41:20] haven't tried, you know, GPT-52 in Codex [41:20] haven't tried, you know, GPT-52 in Codex or in Cursor, it's worth trying. Even [41:22] or in Cursor, it's worth trying. Even [41:22] or in Cursor, it's worth trying. Even for non-technical people, it's just [41:24] for non-technical people, it's just [41:24] for non-technical people, it's just amazing. I think almost being technical [41:26] amazing. I think almost being technical [41:26] amazing. I think almost being technical is sort of a constraint cuz you have a [41:28] is sort of a constraint cuz you have a [41:28] is sort of a constraint cuz you have a pre-existing idea for what these models [41:30] pre-existing idea for what these models [41:30] pre-existing idea for what these models can do and they can do a lot more and [41:32] can do and they can do a lot more and [41:32] can do and they can do a lot more and I'm hearing increasingly about people [41:33] I'm hearing increasingly about people [41:33] I'm hearing increasingly about people doing knowledge work and writing essays [41:35] doing knowledge work and writing essays [41:35] doing knowledge work and writing essays in Cursor instead of just writing code. [41:37] in Cursor instead of just writing code. [41:37] in Cursor instead of just writing code. Wow. Just one thing I'm going to do at [41:39] Wow. Just one thing I'm going to do at [41:39] Wow. Just one thing I'm going to do at the end year year end that just to plug [41:41] the end year year end that just to plug [41:41] the end year year end that just to plug in like a popular trend I see on TikTok [41:44] in like a popular trend I see on TikTok [41:44] in like a popular trend I see on TikTok where there are people who said, "What [41:45] where there are people who said, "What [41:45] where there are people who said, "What is a most unhinged thing I said this [41:48] is a most unhinged thing I said this [41:48] is a most unhinged thing I said this year?" [41:49] year?" [41:49] year?" And it actually does a review of of of [41:52] And it actually does a review of of of [41:52] And it actually does a review of of of all the things that you said. But I [41:54] all the things that you said. But I [41:54] all the things that you said. But I think similarly it'll be a good thing [41:56] think similarly it'll be a good thing [41:56] think similarly it'll be a good thing I'm going to do this at the year end. [41:59] I'm going to do this at the year end. [41:59] I'm going to do this at the year end. Tell me how to live a better life next [42:00] Tell me how to live a better life next [42:00] Tell me how to live a better life next year. Yeah. Give me give me actual [42:03] year. Yeah. Give me give me actual [42:03] year. Yeah. Give me give me actual unvarnished [42:04] unvarnished [42:04] unvarnished uh, opinions and and some direction that [42:06] uh, opinions and and some direction that [42:07] uh, opinions and and some direction that I think it'll be helpful. [42:09] I think it'll be helpful. [42:09] I think it'll be helpful. I love that idea. I'm going for a worse [42:10] I love that idea. I'm going for a worse [42:10] I love that idea. I'm going for a worse life next year. [42:12] life next year. [42:12] life next year. >> [laughter] [42:13] >> [laughter] [42:13] >> [laughter] >> Fantastic. Let's go full degen, guys. [42:16] >> Fantastic. Let's go full degen, guys. [42:16] >> Fantastic. Let's go full degen, guys. Any closing thoughts? [42:18] Any closing thoughts? [42:18] Any closing thoughts? That was a good [42:19] That was a good [42:19] That was a good one. I mean, [42:20] one. I mean, [42:20] one. I mean, the obvious one is we are very actively [42:22] the obvious one is we are very actively [42:22] the obvious one is we are very actively investing in consumer companies and and [42:24] investing in consumer companies and and [42:24] investing in consumer companies and and I genuine I think a lot of people say [42:25] I genuine I think a lot of people say [42:25] I genuine I think a lot of people say this. I genuinely believe that the [42:28] this. I genuinely believe that the [42:28] this. I genuinely believe that the models have gotten to the level of [42:29] models have gotten to the level of [42:29] models have gotten to the level of quality that you can build a real [42:32] quality that you can build a real [42:32] quality that you can build a real scalable app on top of them. Wabi is a [42:34] scalable app on top of them. Wabi is a [42:34] scalable app on top of them. Wabi is a great example of this. [42:36] great example of this. [42:36] great example of this. Um, and so the hope is 2026 will be a [42:38] Um, and so the hope is 2026 will be a [42:38] Um, and so the hope is 2026 will be a huge year for consumer builders, um, not [42:41] huge year for consumer builders, um, not [42:41] huge year for consumer builders, um, not just like consumers as, um, [42:45] just like consumers as, um, [42:45] just like consumers as, um, like consumers being consumers of a [42:47] like consumers being consumers of a [42:47] like consumers being consumers of a product. [42:47] product. [42:47] product. >> Yeah. [42:48] >> Yeah. [42:48] >> Yeah. >> Yes. Well, thank you all for a super fun [42:49] >> Yes. Well, thank you all for a super fun [42:49] >> Yes. Well, thank you all for a super fun year in consumer and AI. We'll be back [42:51] year in consumer and AI. We'll be back [42:51] year in consumer and AI. We'll be back with more next year and Merry Christmas, [42:53] with more next year and Merry Christmas, [42:53] with more next year and Merry Christmas, guys. This is a wrap. Yeah. Happy [42:55] guys. This is a wrap. Yeah. Happy [42:55] guys. This is a wrap. Yeah. Happy holidays. Happy holidays. [laughter] [42:56] holidays. Happy holidays. [laughter] [42:56] holidays. Happy holidays. [laughter] Happy holidays.