AI Rundown: Google Drops Gemini Flash Trio as 3.5 Pro Stalls, Anthropic Reaches $1.5B Copyright Deal, and Poolside Debuts Laguna S 2.1

AI Rundown: Google Drops Gemini Flash Trio as 3.5 Pro Stalls, Anthropic Reaches $1.5B Copyright Deal, and Poolside Debuts Laguna S 2.1 The Rundown highlights Google's latest Gemini efficiency models alongside the ongoing absence of Gemini 3.5 Pro, a landmark $1.5 billion copyright settlement for Anthropic, and Poolside's release of an open-weights coding model. Google launched Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, while Gemini 3.5 Pro remains delayed in partner testing. Anthropic received court approval for a $1.5 billion settlement covering 482,000 pirated book titles, maintaining AI training as fair use. Poolside released Laguna S 2.1, an open-weights coding model using 8B of 118B active parameters that can run on a single desktop. Industry quick hits include Jack Dorsey's workspace Buzz, Substack's AI detector partnership, Anthropic's workflow recording tool, and Sakana AI's Fugu-Cyber. Google Expands Gemini Flash Family While 3.5 Pro Remains Missing Google launched three new models: Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. The release prioritizes efficiency over raw intelligence; 3.6 Flash showed no upgrade over version 3.5 on Artificial Analysis' Intelligence Index and trails rivals like Grok 4.5 and GPT-5.6 Luna. Meanwhile, Gemini 3.5 Pro remains in testing with partners after several delays, and Google announced it has begun a pre-training run for Gemini 4. Anthropic Settles Authors' Copyright Lawsuit for $1.5 Billion A court approved Anthropic's $1.5 billion settlement with book authors over 482,000 works obtained from piracy hubs, equating to nearly $3,000 per title. The agreement follows a 2025 split ruling by Judge William Alsup, which affirmed AI training as legal fair use while holding the pirated 7-million-book stockpile liable. The settlement prevented a jury trial that carried potential damages reaching hundreds of billions of dollars. Poolside Launches Laguna S 2.1 Open-Weights Coder Poolside released Laguna S 2.1, an open-weights coding model trained in under nine weeks. Operating with 8B of 118B parameters at a time and a 1M-token context, it can run on an Nvidia DGX Spark desktop. While it leads U.S. open systems on benchmarks, it continues to trail Chinese open models such as Kimi K3. Industry Quick Hits Jack Dorsey launched early access for Buzz, an open-source workspace featuring AI agent coworkers. Substack integrated Pangram's AI writing detector, Anthropic introduced a 'Record a skill' workflow automation feature for Claude, and Sakana AI unveiled Fugu-Cyber, a multi-agent cybersecurity model.

谷歌更新Gemini Flash系列但Pro仍缺席，Anthropic获批15亿美元版权和解 本期简报汇总了AI领域的最新动态：谷歌推出了主打效率的三款Flash系列模型，但受期待的3.5 Pro仍未推出；Anthropic就涉盗版训练的图书诉讼获批15亿美元和解协议；Poolside推出开源代码模型Laguna S 2.1。 谷歌推出Gemini 3.6 Flash、3.5 Flash-Lite和3.5 Flash Cyber，而备受延期的3.5 Pro仍在合作伙伴测试阶段，同时Gemini 4已启动预训练。 法院批准Anthropic与图书作者达成15亿美元和解协议，针对盗版网站提取的48.2万部作品提供单部近3000美元的赔偿，同时维持AI训练构成合理使用的裁决。 Poolside发布开源权重代码模型Laguna S 2.1，以80亿活跃参数运行于单台桌面设备，在欧美开源榜单领先但仍落后于Kimi K3等中国开源模型。 行业动态涵盖Jack Dorsey推出AI协同办公空间Buzz、Substack集成AI写作检测工具Pangram，以及Sakana AI发布安全模型Fugu-Cyber。 谷歌推出多款Flash模型，Pro旗舰型号继续推迟 谷歌发布了三款Gemini模型，包括3.6 Flash、专注于速度的3.5 Flash-Lite以及用于网络安全的3.5 Flash Cyber，主打运行效率。在测试中，3.6 Flash落后于同等价位的Grok 4.5与GPT-5.6 Luna，在Artificial Analysis智能指数上相比3.5并未显著提升。官方称3.5 Pro仍在与伙伴进行测试，而Gemini 4的预训练工作也已启动。 Anthropic获批15亿美元图书版权和解案 美国法院批准了Anthropic针对作家版权诉讼的15亿美元和解方案。该案始于2024年，指控Claude利用盗版平台文本进行训练。主审法官曾于2025年判定AI模型训练属于合理使用，但将储存盗版图书库视作侵权。和解协议覆盖48.2万部作品，平均单本赔偿近3000美元，避免了原定可能带来数千亿美元索赔的陪审团审判。 开源模型进展与行业速递 Poolside推出了开源代码模型Laguna S 2.1，采用总计1180亿参数、单次激活80亿参数的设计，可在单台Nvidia DGX Spark上运行。其他动态中，Jack Dorsey上线了让AI智能体作为同事参与群聊的Buzz早期预览版，Substack与Pangram合作内置AI写作检测，Anthropic推出了支持录屏生成工作流技能的Record a skill功能，Sakana AI发布了网络安全多智能体模型Fugu-Cyber。

[Read Online]({{live_url}}) ^| [Sign Up](https://www.therundown.ai/subscribe?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) ^| [Advertise](https://therundownai.typeform.com/to/kraZ1TSO?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) **Good morning, {{ first\_name | AI enthusiasts }}.** Google just added three more models to the Gemini family. Somehow, that made the missing one even harder to ignore. A new Flash trio covers speed, low-cost use, and cybersecurity, and the company’s “most ambitious pre-training run yet” for Gemini 4 is already in progress. But the delayed 3.5 Pro model expected to challenge frontier rivals? Still nowhere to be found. **Reminder** *— Our next live workshop is today at 12 PM EST! Join and learn how to go from idea to working product without a technical team using AI tools for research, prototyping, iteration, and more. RSVP here.* **In today’s AI rundown:** Google drops new Flash models as Pro stays missing AI's biggest copyright bill comes due Sell high-value AI workflow audit as a consultant Poolside resurfaces with a top open-weights coder 4 new AI tools, community workflows, and more **LATEST DEVELOPMENTS** Image source: Google **The Rundown:** Google just [released](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) a trio of Gemini models, led by 3.6 Flash, alongside a speed-focused 3.5 Flash-Lite and a security-fine-tuned 3.5 Flash Cyber, prioritizing efficiency over intelligence, as the long-awaited 3.5 Pro remains in testing. **The details:** 3.6 Flash brings efficiency upgrades over its predecessor, but is still topped by similar-priced rivals like Grok 4.5 and GPT-5.6 Luna on a variety of tests. On Artificial Analysis’ Intelligence Index, 3.6 Flash strangely shows no upgrade over 3.5, slotting in near Meta’s Muse Spark and Z AI’s open GLM 5.2 model. 3.5 Pro remains unreleased after several delays, with Google's Logan Kilpatrick saying the model is "testing with partners" and "will hopefully land soon." Google also ended the announcement with an update on Gemini 4, saying the company has kicked off its "most ambitious pre-training run yet". **Why it matters:** While a lot of people use Gemini models, which makes efficiency upgrades useful, these scores and the absence of a strong frontier model only add to the perception that Google is falling behind (with even Elon [chiming](https://x.com/elonmusk/status/2079597978076291301?s=20&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) in to take a jab). A lot is riding on the company’s 3.5 Pro and its 4 class that is now in training. **The Rundown:** Unwrap is a customer intelligence platform that pulls your feedback (surveys, reviews, support tickets, social comments) into one view, using AI and NLP to surface actionable insights and deliver them straight to your inbox. **With Unwrap, you get:** All customer feedback automatically categorized Query feedback using Unwrap Assistant, or in your favorite tools using Unwrap's MCP Real-time alerts and a clear view of customer sentiment as it shifts The tools trusted at scale by Perplexity, DoorDash, Southwest, lululemon, and Oura Unwrap is offering a trial of its tools to Rundown AI subscribers! Just [grab 20 minutes with the team to get set up](https://cal.com/unwrap/quick-connect?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole). Image source: Images 2.0 / The Rundown **The Rundown:** Anthropic [secured](https://www.reuters.com/world/us-judge-approves-anthropics-15-billion-settlement-copyright-lawsuit-2026-07-20/?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) court approval for a $1.5B settlement with book authors, clearing payouts of nearly $3,000 a title across 482K works pulled from piracy sites — a record for U.S. copyright cases that leaves AI training standing as fair use. The case dates back to 2024, when novelist Andrea Bartz and other writers accused Anthropic of training Claude on texts lifted from internet piracy hubs. Judge William Alsup split the case in 2025, ruling AI training as legal fair use but treating Anthropic's 7M-book pirated stockpile as its own violation. The settlement spared Anthropic a jury trial that was set for last December, where it faced potential damages of several hundred billion (!) dollars. **Why it matters:** $1.5B sounds like a lot until you weigh it against a trial with potential damages in the hundreds of billions, and a fair-use ruling that now goes untouched while Anthropic keeps training on (legally purchased) books. With other labs facing similar suits, $3,000 a work may be the new rate for making copyright issues disappear. **The Rundown:** In this guide, you will learn how to turn your AI knowledge into a focused consulting offer. Using Claude, you will choose a narrow problem, collect client context, and create a one-page opportunity brief leading to an implementation project. **Step-by-step:** Ask Claude to interview you about the work, industries, and tools you know best, then suggest three consulting angles. Pick the one that works best Ask Claude to build an AI assessment for a prospect to uncover AI opportunities in their business, scored out of 100, and publish as a live Artifact Save their answers. Then, feed those responses, along with the prospect’s website, into Claude, asking it to find the highest-return workflow to fix with AI Ask Claude to turn the findings into a client-ready slide deck, complete their audit result, the bottleneck, practical solutions, and a sensible next step **Pro tip:** Before the audit, have Claude research the best tools, templates, and examples for your niche, and keep that library ready to match against bottlenecks. **[Get resources, prompts, and a full step-by-step video guide here](https://app.therundown.ai/guides/sell-a-high-value-ai-workflow-audit-as-a-consultant?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole)**. **The Rundown:** Bridging the gap between a multimodal sandbox demo and a scalable product is a massive engineering challenge. The Startup technical guide on generative media provides the architectural blueprint to actually build and deploy your ideas. **Inside the guide, you’ll learn how to:** Architect multimodal workflows that orchestrate media pipelines Turn models like Veo and Lyria into high-traffic, user-facing apps Build backends for economic scaling, deterministic output, and reliability Stop experimenting and start building. [Download the guide](https://cloud.google.com/resources/content/technical-guide-genmedia?utm_source=gfs&utm_medium=embedded&utm_campaign=FY26-Q3-GLOBAL-GCP41128-website-dl-startupgenmedia-195260&utm_content=Rundown_July_2) to get started. Image source: Poolside **The Rundown:** Poolside [launched](https://poolside.ai/blog/introducing-laguna-s-2-1?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Laguna S 2.1, a new open-weights coding model that tops the Western open-source field while staying small enough to live on a single desktop — the strongest sign of America’s response to China's open-model lead. S 2.1 uses just 8B of 118B parameters at a time, letting it fit on one Nvidia DGX Spark desktop box, with the weights free to grab on Hugging Face. Poolside’s model tops coding benchmarks for U.S. open systems, but still trails significantly behind Kimi K3 and several other Chinese open models. Laguna arrives a week after Mira Murati's TML [released](https://www.therundown.ai/p/openai-new-230-ai-agent-control-pad?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) its Inkling as its first-ever model, with both expanding the U.S.’ limited open source pool. Poolside said training took under nine weeks, with co-CEO Jason Warner [telling](https://www.forbes.com/sites/iainmartin/2026/07/21/american-open-source-labs-think-they-can-beat-chinas-best-ai-startups/?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Forbes the aim is to create "the most capable open model in the West." **Why it matters:** TML's Inkling made self-hosted coding AI more interesting for U.S. buyers last week, and Laguna now adds a smaller option with even stronger results. Two releases don’t erase China's deep lead (and Kimi is still lapping the field), but they at least show that American open-source has momentum instead of just a gap. **QUICK HITS** 🔒 [Incogni](https://deal.incogni.io/aff_c?offer_id=6&aff_id=1889&url_id=1&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) - remove your personal data from the web so scammers and identity thieves can’t access it. Use code **RUNDOWN** to get 55% off\* 🚀 [Laguna S 2.1](https://www.rundown.ai/tools/laguna-s-2-1?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) - Poolside's open-weight coding AI with a 1M-token context ⚡️ [Gemini 3.6 Flash](https://deepmind.google/models/gemini/flash/?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) - Google’s newest Flash variant with upgraded efficiency 🎆 [Qwen-Image-3](https://www.rundown.ai/tools/qwen-image-3-0?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) - Alibaba’s new image model with strong text rendering *\*Sponsored Listing* **Mozilla** released its 2026 State of Open Source AI Report last week, revealing open models now rival Big Tech's best. [Learn more](https://stateofopensource.ai/?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole).\* **Twitter and Block founder Jack Dorsey** [unveiled](https://x.com/jack/status/2079605800998146171?s=20&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Buzz in early access, an open-source workspace where AI agents join team chats as coworkers. **Substack** [introduced](https://post.substack.com/p/against-claudefishing?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) a new partnership with Pangram that embeds its AI writing detector tool into the platform for both readers and publishers. **Anthropic** [rolled out](https://x.com/claudeai/status/2079595988998554047?s=20&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Record a skill, a feature (similar to Codex’s [upgrade](https://x.com/OpenAIDevs/status/2067681320281723113?s=20&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole)) that allows users to screen-record a workflow to let Claude create a skill around the task. **Sakana AI** [released](https://sakana.ai/fugu-cyber-release/?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Fugu-Cyber, a security model that bundles specialist agents behind one API, claiming benchmarks on par with GPT-5.5-Cyber and Mythos Preview. **COMMUNITY** Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier. Today’s workflow comes from reader **Jonathan** in Waxhaw, NC: “I'm writing a novel. Asking AI 'is it good?' receives flattering positive feedback — annoying. I've developed an evaluation protocol which prompts multiple AIs to assess the material through market research simulation, from four perspectives: reader acceptance, publication acceptance, production acceptance, and marketability. The protocol instructs the AI to generate four personas — some antagonistic, some YA-biased, some literature aficionados, and some business-biased. I've input prose and pitch decks, and results have identified gaps, inconsistencies, and positives. Running across multiple AIs gives results that differ but frequently identify the same strengths and weaknesses. These results are LLM training-biased, but different draft revisions and comparing across AIs gives valuable feedback.” How do you use AI? Tell us [here](https://www.videoask.com/ffohxg3vc?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole). Read our last AI newsletter: [Claude disproves an 87-year-old math problem](https://www.therundown.ai/p/claude-disproves-an-87-year-old-math-problem?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Read our last Tech newsletter: [China’s plan to knock an asteroid off course](https://tech.therundown.ai/p/china-bold-plan-to-knock-an-asteroid-off-course?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Read our last Robotics newsletter: [The world’s first humanoid cage fight](https://robotnews.therundown.ai/p/world-first-humanoid-cage-fight?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) Today’s AI tool guide: [Sell a high-value AI workflow audit as a consultant](https://app.therundown.ai/guides/sell-a-high-value-ai-workflow-audit-as-a-consultant?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) RSVP to next workshop today @ 12 PM EST: [No-code product building](https://app.therundown.ai/live/no-code-product-building-from-idea-to-product-market-fit-with-ai?utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=google-s-gemini-lineup-has-a-pro-sized-hole) See you soon, *Rowan, Zach, Shubham, and Jennifer — the humans behind The Rundown*