Marketers Fake AI Prompts for Boss Approval as Feeds Flood With Slop

Bloomberg

As corporate managers outsource creative judgment to chatbots and synthetic slop crowds public feeds, social media is abandoning personal sharing for broadcast entertainment while authentic connection retreats into private direct messages.

AI-assisted summary of AI Is Upending the Lives of People Who Do Social Media Professionally — original source. Check the original for full context.

Source published: 2026-09-24

Key ideas

Executive adoption of generative AI has triggered widespread 'AI brain' in corporate marketing, where managers offload creative judgment onto chatbots without formal training or guidelines, deeply demoralizing creative staff. (5:00)
Instructional and utility-based content creators face rapid obsolescence from frictionless LLM troubleshooting, whereas creators offering parasocial connection and idiosyncratic points of view retain a defensible moat. (10:00)
Spam accounts are exploiting platform discovery algorithms by publishing dozens of AI-generated clips daily, with multiple top-viewed Instagram Reels now relying on synthetic footage that monetizes viewer confusion. (22:45)
Public social media feeds have fully transitioned into broadcast entertainment optimized for non-follower discovery, while genuine interpersonal sociality has migrated almost entirely into private direct messages. (30:00)

Detailed notes

## Background Published by Bloomberg on September 24, 2026, via podcast RSS, this live episode of Odd Lots was recorded at The Vermont in Hollywood. According to the show description, it provides a primer on social media's AI era featuring commentators Rachel Karten (author of the Link in Bio newsletter) and Taylor Lorenz (founder of User Mag) exploring feed uncanny valleys, AI slop, changing creative roles, and whether internet time is shifting toward chatbots.

## TL;DR Odd Lots hosts Tracy Alloway and Joe Weisenthal convene internet culture experts Rachel Karten and Taylor Lorenz to analyze how generative AI is upending the creator economy and corporate marketing. The panel details the demoralizing rise of 'AI Brain' management where executives outsource creative taste to chatbots, alongside audience double standards that tolerate viral AI fruit slop while penalizing commercial brands that cut corners. They evaluate practical agency use cases versus copywriting failures, push back against apocalyptic AI existential risk narratives, and debate the shift toward ambient computing. Ultimately, the discussion illuminates how feeds are fracturing into algorithmic entertainment broadcasts driven by non-follower reach, forcing genuine human connection into private DMs and group chats.

## Atmosphere & style Recorded before an energetic theater audience in Los Angeles, the conversation carries a lively, fast-paced, and sharp conversational rhythm fueled by industry cynicism. The dynamic between the hosts and panelists blends witty, sardonic humor with insider clarity, as deadpan observations about viral AI Mentos trucks and corporate Copilot mockups draw genuine amusement from the crowd. It feels less like an academic technology lecture and more like a cathartic, behind-the-scenes trade debrief where seasoned digital veterans bluntly diagnose the uncanny, exhausting realities of modern internet work.

## Bulletin - `05:00` Executive adoption of generative AI has triggered widespread 'AI brain' in corporate marketing, where managers offload creative judgment onto chatbots without formal training or guidelines, deeply demoralizing creative staff. - `10:00` Instructional and utility-based content creators face rapid obsolescence from frictionless LLM troubleshooting, whereas creators offering parasocial connection and idiosyncratic points of view retain a defensible moat. - `22:45` Spam accounts are exploiting platform discovery algorithms by publishing dozens of AI-generated clips daily, with multiple top-viewed Instagram Reels now relying on synthetic footage that monetizes viewer confusion. - `30:00` Public social media feeds have fully transitioned into broadcast entertainment optimized for non-follower discovery, while genuine interpersonal sociality has migrated almost entirely into private direct messages.

## Key takeaways - `00:00` The advent of virtually free and frictionless generative content has induced existential anxiety and negative vibes across the influencer and social media marketing sectors. - `05:00` Corporate managers increasingly outsource creative instincts to LLMs—such as asking Claude if a joke is funny—forcing creative staff to fake AI-branded concepts just to get client approval. - `05:00` Consumers view creative production quality as a proxy for overall operational integrity, causing audiences to distrust physical brands that cut corners with automated, synthetic advertising. - `10:00` Audiences exhibit a double standard toward AI media: they binge low-stakes synthetic entertainment like fruit soap operas, but react aggressively when commercial brands deceive them. - `10:00` Informational DIY and instructional creators are most vulnerable to displacement by AI chatbots, while personality-driven and commentary-based creators maintain human defensibility. - `15:00` The political opposition to AI remains fragmented because local environmental coalitions fighting physical data centers have fundamentally divergent motivations from existential risk theorists. - `15:00` Anthropomorphizing software models by assigning them human intent or intellect distorts regulatory clarity, corporate accountability, and legal liability. - `17:15` In professional marketing, AI serves effectively as an administrative utility and visual pitch deck mockup tool, while generative LLM copywriting consistently produces bland and ineffective output. - `17:15` Media consumption is fundamentally constrained by a daily waking ceiling of around 16 hours, making conversational chatbots a potential threat to traditional feed screen time. - `20:00` The shift toward ambient computing, wearables, and voice interfaces may reduce passive smartphone screen scrolling, though it risks creating alternative forms of social isolation. - `22:45` Automated accounts brute-force platform distribution by publishing up to 50 synthetic videos daily, monetizing user confusion by disguising AI simulations as behind-the-scenes film footage. - `25:09` Meta's Threads has fostered an intensely toxic environment due to recommendation algorithms surfacing out-of-context replies, amplified by hallucinating AI news summaries. - `25:09` Generic lifestyle influencer content is largely dead; sustainable audience growth in modern feeds requires serialized, episodic formats resembling television shifts. - `25:09` Major talent agencies like CAA are constructing likeness vaults to archive client voices and digital replicas for authorized synthetic licensing. - `30:00` Instagram distribution mechanics explicitly prioritize video reels and carousels for non-follower reach, while static feed images serve almost entirely for legacy follower retention. - `30:00` Interpersonal social media has structurally split: open feeds function as broadcast television for strangers, while actual relationship maintenance occurs privately through direct messages.

## Detailed walkthrough ### `00:00` Live in Hollywood: The Odd Lots Podcast Tackles the Creator Economy

The episode opens with the *Odd Lots* podcast hosting its first-ever live show in Los Angeles at the Vermont in Hollywood. Hosts Tracy Alloway and Joe Weisenthal note that being in LA demands confronting the massive, often invisible-to-traditional-media world of the influencer and creator economy. With creators amassing millions of followers on platforms like Instagram without mainstream notoriety, the hosts explore the existential crisis currently facing digital workers: what happens to creators and social media professionals when generating content becomes virtually free and effortless through generative AI?

To unpack these questions, Tracy and Joe bring out two leading commentators on internet culture and social media strategy: - **Taylor Lorenz**, technology journalist and founder of the Substack publication *User Mag*, who covers online culture, creator dynamics, tech, and policy. - **Rachel Karten**, social media consultant and author of the Substack newsletter *Link in Bio*, regarded as the premier trade publication for corporate social media practitioners and brand marketers.

When Joe kicks off the conversation by asking about the current "vibes" across social media, the immediate, blunt response from the stage is unanimous: "Terrible."

`05:00` Exploding Coke Trucks and "AI Brain": How Generative Tools Are Warping Marketing Teams

Highlighting the bizarre texture of modern online feeds, Lorenz shares an anecdote from backstage just before coming out. She admits she was briefly fooled by an absurd, hyper-viral AI video: a Coca-Cola truck overturned and exploded fountainous soda while someone tried to scoop it with a bucket, only for the next shot to reveal a crashing Mentos truck next door, launching a bystander rocketing through the sky. The caption read, "the AI videos that your mom sends you," prompting Weisenthal to joke that older internet users remember when Mentos-and-Diet-Coke videos were the only content that existed on the entire web.

Turning to the professional workplace, Weisenthal asks whether social media managers view AI as an empowering productivity augmenter or as an outright replacement. Karten explains that for marketing staff, AI is primarily arriving as an unwelcome mandate imposed from executive management. In her viral piece titled *"Does Your Boss Have AI Brain?"*, she interviewed around 100 marketers to document how managers are completely offloading their human creative judgment onto large language models:

- One social media manager spent hours crafting a nuanced, funny meme, only for their boss to feed it into Anthropic's Claude and literally ask the chatbot: *"Is this funny?"* - Another marketer discovered their supervisor would only greenlight creative concepts if they appeared to originate from Microsoft Copilot; as a result, the employee began mocking up their own original ideas inside a fake Copilot interface just to get approval. - An interviewee summarized the prevailing corporate atmosphere as "demoralizing as hell."

Karten contrasts this chaotic deployment with her past tenure at Condé Nast. When the publisher onboarded tools like Asana, the organization spent months carefully training staff and developing structured workflows. By contrast, AI adoption in marketing has been implemented with no guardrails, no shared philosophy, and zero formal training, leaving creative workers feeling stripped of their agency.

```mermaid flowchart TD accTitle: Boss AI Brain Workflow Loop accDescr: Diagram illustrating how marketing staff must bypass or appease managerial reliance on AI tools.

subgraph GroundReality ["Creative Social Media Manager"] A["Drafts organic meme or campaign"] end

subgraph ExecutiveIntervention ["'AI Brain' Management"] B{"Manager outsources gut check"} C["Asks Claude: 'Is this funny?'"] D["Only trusts ideas branded by Copilot"] end

subgraph WorkerWorkarounds ["Employee Counter-Strategy"] E["Worker fakes Copilot UI mockup"] F["Idea gets signed off"] end

A --> B B --> C B --> D D --> E E --> F ```

Addressing the pushback over whether consumers will eventually stop caring about AI-generated marketing, Karten argues that brand creative has always functioned as a proxy for operational integrity. When iconic brands like Nike invest deeply in craft, consumers interpret that effort as proof of quality. If a beloved brand—such as Chipotle or any fast-food chain—resorts to automated, synthetic imagery, audiences immediately question what shortcuts the company is taking with its physical ingredients and core product.

Lorenz adds that while consumers claim to hate AI, their behavior tells a contradictory story. On YouTube, creators routinely read AI-generated scripts to glowing comment sections praising their "brilliant analysis." Audiences generally do not mind AI content when they find it genuinely entertaining; their hostility surfaces when they feel actively duped or tricked. The panel points to the rise of bizarre viral formats like "fruit slop" and "fruit soap" (including automated soap operas like *Fruit Love Island*, depicting an orange falling in love with a carrot): users will rage against synthetic content in one post's comments, only to binge-watch AI-generated anthropomorphic produce in the very next swipe.

`10:00` Low Stakes Memes, Incumbents, and the Threat to "How-To" Content

The panel digs into the double standard audiences hold when evaluating AI-generated material. Viewers navigate a weird paradox: they give low-stakes, throwaway humor—like an AI-generated meme of fruit mimicking *Love Island*—an immediate pass because they do not care about the craft of a casual meme. Yet if a commercial brand posts an AI-generated menu or promotional asset, consumers react with fury. In short, the exact same person laughing at and texting "AI fruit slop" to their friends might simultaneously excoriate a corporation online for cutting human corners.

This tension prompts a deeper question about the broader creator ecosystem: in an AI-saturated world, does leverage accrue to incumbent creators who have already built personal trust, or will plummeting production costs spark another wave of democratization where anyone churns out content for free? The answer hinges entirely on format and genre:

```mermaid flowchart TD accTitle: Impact of AI on Creator Archetypes accDescr: Flowchart contrasting how AI affects utility-driven how-to creators versus personal and opinion-driven creators.

subgraph Utility["Utility-Driven Content"] U1["'How-To' & Informational Content (Dishwasher repair, beauty tutorials)"] U2["Easily Replaced by LLMs (ChatGPT, quick troubleshooting)"] U1 --> U2 end

subgraph Parasocial["Parasocial & Point-of-View Content"] P1["Personal Life, Commentary & Point of View"] P2["High Moat & Authenticity (Difficult to automate or replicate)"] P1 --> P2 end ```

Creators who rely purely on utility and instruction are the most vulnerable. Early social media democratization was heavily built on informational "how-to" formats—such as the earliest beauty and makeup tutorials on YouTube. Today, AI models are rapidly becoming superior, frictionless reference manuals. As an everyday example, one speaker shares that her own mother proudly used ChatGPT to troubleshoot and repair a broken toilet—a DIY feat she never would have expected her to pull off on her own. While utility-driven creators face obsolescence, personalities who offer raw parasocial connection, idiosyncratic opinions, or non-replicable commentary will maintain their moat.

`11:30` Mid-Roll Break: Agentic Commerce and Workplace AI

The episode cuts to sponsor messages highlighting how artificial intelligence is actively reshaping enterprise operations and commercial infrastructure: - **PayPal Open**: Explores the psychological friction in digital shopping—the split-second hesitation before hitting "Buy Now"—and outlines how PayPal is positioning its 25-year payments infrastructure for "agentic commerce," allowing businesses to deploy AI-driven discovery while retaining brand trust and customer relationships. - **ChatGPT Work**: Pitching beyond basic search engine queries toward deep task execution, highlighting Work Mode (available across Plus and Pro tiers) to coordinate across local files and apps, grinding unstructured source materials and briefs into reviewable first drafts. - **LinkedIn Advertising**: Marketers in "Dashboard Confessions" call out vanity reach metrics ("bullspend") and prioritize revenue over surface-level impressions, offering a $250 credit match. - **Dupixent (dupilumab)**: Prescription medical disclosure for adults suffering from moderate-to-severe eczema, targeting systemic inflammation sources with detailed warnings regarding potential allergic reactions, eye issues, and joint pains.

`15:00` The AI Discourse Cycle: Data Centers, Existential Risk, and Anthropomorphism

Returning to the live stage, the conversation shifts to the dizzying churn of online discourse. Internet controversy moves so fast that debates seem to expire overnight; what felt like an urgent public crisis a month ago is suddenly dismissed as old news:

> "Oh, you're still arguing about data centers? Wasn't that a July issue? That was so July."

Taylor Lorenz observes that while political consultants like Morris Cafritz argue local fights over data center infrastructure and energy grids are primed to dominate the 2026 midterm election cycle rather than 2028, public outrage shifts continuously. A key political divide exists between anti-AI camps: the environmental and civic critics upset about physical infrastructure have entirely different motives and values from the "X-risk" (existential risk) camp fixated on runaway machine intelligence. Because their goals are so misaligned, they have completely failed to build a united front.

Addressing the existential risk narrative directly, Lorenz pushes back on the panic: "First of all, it's not going to kill us. It's actually so hard to kill all humans on Earth, even if we wanted to." Drawing on a quip from a former FTC commissioner—who joked that humanity survived the Black Death when it wiped out half the population—the hosts emphasize that deadliness does not equal superior intelligence; a biological virus can decimate populations without possessing an intellect.

The real danger lies in how easily people anthropomorphize computer code: - Assigning human agency, emotions, and moral intent to software skews clear thinking. - While anthropomorphic metaphors help lay audiences grasp technical mechanics, treating software programs like living entities muddies critical policy, corporate accountability, and legal liability.

`17:15` Practical Uses in Marketing: Pitch Deck Mockups vs. Copywriting

The discussion turns to whether marketers are actually getting practical value out of AI tools, or if the technology merely produces internal nausea—such as an agency employee pitching a sharp joke only to be told by a superior, "Well, my buddy Claude doesn't think that it's funny."

Rachel Karten notes that AI's most effective role in agency environments right now is visual translation. Strategists use image and concept generators inside pitch decks to build quick mockups for executives or clients who lack the imagination to visualize an abstract idea. AI bridges the gap between a written concept and a physical campaign, helping sell a real-world activation that will ultimately be executed by human craftspeople. Conversely, both panelists express deep skepticism toward using LLMs for generative copywriting, social media post ideation, or brand captions, warning that it tends to yield bland, ineffective results.

Lorenz details her personal workflow, cautioning humorously, "Don't cancel me because people get so mad when you talk about it." She refuses to use AI for actual prose writing because expressing her own thoughts is the entire point of the craft. Instead, she treats AI strictly as an administrative utility: - Drafting basic SEO-driven descriptions and mechanical metadata - Resizing, extending, and manipulating image canvases in design tools like Canva - Handling repetitive operational tasks to save creative energy for human reporting

The segment closes by opening up a broader structural bottleneck: human attention. Even if generative tools flood feeds with limitless media, human capacity remains capped at 24 hours in a day—meaning that after eight hours of sleep, platforms are ultimately fighting over a finite ceiling of roughly 16 hours of daily scrolling.

`20:00` Chatbots vs. Feeds, Screen Time Panic, and the Ambient Computing Debate

The conversation turns to how people allocate their digital attention, raising the question of whether conversational AI is cannibalizing passive social media scrolling to the detriment of marketers and brands. Joe Weisenthal notes a personal dilemma: whenever he posts on social media, he faces an immediate barrage of insults from users saying, "this guy's an idiot" or claiming he has lost his mind. By contrast, a chatbot like Claude or ChatGPT provides a frictionless, pleasant conversation where nobody dunks on him. Taylor Lorenz and Rachel Karten push back on the equivalence, arguing that Twitter exists in a combative league of its own compared to TikTok or Instagram, where users consume feeds like television for pure entertainment. Furthermore, chatting with an AI serves a fundamentally different psychological and utilitarian function than direct-messaging a real friend.

The panel dives into an energetic debate over screen time and hardware form factors. Taylor argues that the widespread moral panic over screen time misses a structural shift: the traditional smartphone slab is destined to decline rapidly. In her view, people are increasingly shifting toward ambient computing, relying on earbuds, smart glasses, connected home appliances, and wearable voice accessories (like recently unveiled whispering pendant bands) rather than staring blankly at a glass display. If people check a physical screen less frequently, she contends, they will be sucked into algorithmic doomscrolling less often.

Joe and the panel counter with healthy skepticism about whether this ambient hardware future is genuinely liberating or simply substituting one antisocial habit for another. Looking up a tutorial to repair a dishwasher or fix a toilet on a screen previously replaced hiring a plumber and having a genuine face-to-face human interaction. Joe jokes about the absurdity of smart homes creating artificial friction, quipping that soon you will need two-factor authentication just to grab a snack: "open your refrigerator, open up your Google Auth and put in your six-digit code, and then I'm back looking at the screen."

```mermaid flowchart TD accTitle: Ambient Computing vs Screen Feed Consumption accDescr: Diagram contrasting passive phone feed consumption with conversational AI and ambient hardware.

subgraph Feeds ["Traditional Feed Browsing"] A["Instagram & TikTok: Passive Entertainment / TV"] B["Twitter / X: Public Discourse & Conflict"] end

subgraph Chatbots ["Conversational & Ambient Alternatives"] C["AI Chatbots (Claude / ChatGPT): Safe, Frictionless Interaction"] D["Ambient Hardware: Smart Glasses, Wearables & Appliances"] end

subgraph Impact ["Audience Shift"] E["Erosion of Traditional Screen Time"] F["Isolation / Antisocial Friction"] end

A --> E B --> C C --> E D --> F ```

`22:45` AI Slop, Platform Inertia, and the Monetization of Confusion

Tracy Alloway shifts the focus to how major platforms are handling the influx of automated "AI slop" and virtually free synthetic content. This dynamic presents two divergent paths: either disillusioned users flee public feeds entirely in favor of private group chats, DMs, and community spaces like Discord, or platform algorithms must become hyper-curated engines capable of detecting and delivering users' exact tolerance for synthetic media. Taylor Lorenz argues that traditional platforms are currently stuck on their back foot and failing to adapt. She views Meta in particular as an aging 2010s dinosaur that has already passed its cultural peak, leaving Google comparatively better positioned for the next phase of internet architecture.

Rachel Karten highlights the aggressive, short-term tactics currently degrading feed quality. Automated AI accounts exploit algorithmic discovery by posting at inhuman scale—sometimes churning out 50 posts a day—treating Instagram's distribution like a slot machine until one hits the jackpot. Examining recent data on the top 20 most-viewed Instagram Reels of all time, Karten discovered that two or three of the record-holders were recent AI-generated videos. One viral clip mimicked behind-the-scenes movie footage, pretending to show a giant practical-effects wave crashing into houses on a film set. By borrowing visual cues of raw reality and "behind-the-scenes" authenticity, the creators engineered massive engagement purely by confusing the audience about whether what they were watching was real. Karten condemns this practice of "monetizing confusion," warning that while it delivers immediate viral metrics, it inflicts severe long-term damage on platform trust.

- Automated AI spam accounts flood platforms with up to 50 posts per day to brute-force algorithmic virality. - Multiple entries among the top 20 most-viewed Instagram Reels of all time are now completely AI-generated clips. - Viral synthetic videos increasingly masquerade as "proof of reality" or behind-the-scenes filmmaking footage to farm view counts from bewildered viewers.

`25:09` The Toxicity of Threads and the Formula for Modern Follower Growth

The conversation moves to Meta's text-based platform, Threads, with Tracy asking who is actually using it. Rachel admits that her own mother is an avid user, but reveals that she recently published a newsletter declaring Threads to be the single most toxic social network online today. Taylor enthusiastically agrees, claiming that "Twitter is like a wonderful fairyland compared to Threads." Because the Threads recommendation engine frequently drops users into the middle of hostile, out-of-context replies, it cultivates intense mob-like pile-ons and rapid groupthink. Compounding the chaos, Threads features AI-generated news summaries that routinely hallucinate, invent false brand drama, and present outright fabrications as fact.

Turning to actionable advice for creators looking to reach millions of followers in 2026, Rachel and Taylor explain that the traditional model of the generic lifestyle influencer—sharing mundane daily life—is effectively dead and actively provokes audience resentment. To succeed today, creators must treat their channels like serialized episodic television shows with a clear, reliable throughline: - **Serialized "TV Show" Formats**: Audiences respond to tightly themed, episodic content. Karten highlights a restaurant line cook (Gilley Cooks) who wears Meta smart glasses while preparing dinner services, racking up over 500,000 followers in two to three months simply by framing every daily video as a new episode of his shift. - **Likeness Rights and the CAA Vault**: Talent agencies are preparing for synthetic duplication. Taylor notes her reporting for *Vanity Fair* on the "CAA Vault," where Creative Artists Agency archives digital replicas of top stars' voices and likenesses so clients can license AI-generated variations of themselves—prompting Joe to joke that Bloomberg will soon license their voices to produce endless automated episodes of *Odd Lots*. - **Feed Distribution Mechanics**: Sharing guidance Instagram provided to her *Link in Bio* readership, Rachel breaks down the platform's current reach hierarchy: Carousels and Reels are specifically prioritized to be served to non-followers. Because the platform algorithm is entirely fueled by non-follower reach, content must immediately appeal to strangers rather than relying on inside jokes understood only by existing fans.

`30:00` The Architecture of Modern Feeds and the Shift to DM-Based Community

The discussion turns to the segmented mechanics of modern platforms like Instagram, where content formats serve strictly bifurcated purposes. Algorithmic discovery mechanisms now dictate that if a piece of media is "not entertaining enough," the platform simply will not surface or distribute it across the main channel or discovery feed. In this environment, static single-image posts have been relegated to reaching an account's existing legacy audience rather than capturing new reach, while true interpersonal community and authentic connection have migrated entirely into Stories and Direct Messages (DMs).

This structural shift was underscored during a conversation at Bloomberg Screentime, where Instagram head Adam Mosseri observed that the foundational "social" element of social media no longer centers on broadcasting personal milestones, such as sharing photos of one's family to an open feed. Instead, modern sociality occurs almost exclusively in private backchannels through DMing friends external content—such as memes or video clips—that reminded the sender of the recipient. The speakers note that while it feels "very unfortunate" to realize that public feeds have abandoned personal life updates in favor of passive broadcast entertainment, Mosseri’s framing is deeply "clarifying" about what connection has truly become online.

```mermaid flowchart TD accTitle: Instagram Content Segmentation Model accDescr: Diagram illustrating how platforms divide content between broad public entertainment and private social connection.

subgraph BroadReach["Public Algorithmic Broadcast"] FeedReels["Main Channel & Reels"] --> Dist["Algorithmic Reach & Entertainment"] StaticPosts["Single-Image Feed Posts"] --> CoreAudience["Legacy Existing Followers"] end

subgraph PrivateConnection["Private Social Channels"] StoriesDMs["Stories & Direct Messages"] --> ActiveDM["DMs & Meme Sharing"] ActiveDM --> TrueCommunity["Personal Community & Friendship"] end ```

- **Content bifurcation**: Main feed algorithmic surfaces prioritize raw entertainment value over social updates, while single-image static posts serve primarily as retention tools for established followers. - **Adam Mosseri's thesis**: The "social" aspect of social media has transitioned from broadcasting personal life events on public feeds to privately sending DMs and relatable memes to individual friends.

`30:35` Odd Lots Live Wrap-Up, Credits, and Commercial Announcements

Hosts Tracy Alloway and Joe Weisenthal conclude their live Odd Lots recording session at The Vermont in Hollywood, thanking guest commentators Taylor Lorenz (founder of *User Mag*) and Rachel Karten (author of the *Link in Bio* newsletter) for their insights into how generative AI and shifting algorithmic distribution are transforming the creator economy and social media labor.

The hosts direct listeners to their respective social accounts and acknowledge the production team before promoting the daily *Odd Lots* newsletter at Bloomberg.com/oddlots and the 24/7 community Discord server. The broadcast concludes with promotional spots highlighting collegiate football quarterback Jaxson Dart's blanket collaboration with Minky Couture, as well as an announcement for the Bloomberg Tech summit in London covering scale, safety, and sovereignty in AI, presented by Salesforce alongside IDA Ireland and Schneider Electric.

- **Host & guest handles**: Tracy Alloway (`@tracyalloway`), Joe Weisenthal (`@thestalwart`), Taylor Lorenz (`@TaylorLorenz`), and Rachel Karten (`@milkcarton`). - **Production team credits**: Carmen Rodriguez (`@CarmenArman`), Dash Bennett (`@dashbot`), Kel Brooks (`@KaleBrooks`), and Kevin Lozano (`@KevinLloydLozano`). - **Commercial notes**: Ole Miss quarterback Jaxson Dart's partnership collection with blanket company Minky Couture; Bloomberg Tech event in London on November 2–3 examining enterprise AI trade-offs with sponsors Salesforce, IDA Ireland, and Schneider Electric.

## Notable quotes > `05:00` the AI videos that your mom sends you

> `05:00` Is this funny?

> `05:00` demoralizing as hell

> `15:00` Oh, you're still arguing about data centers? Wasn't that a July issue? That was so July.

> `15:00` First of all, it's not going to kill us. It's actually so hard to kill all humans on Earth, even if we wanted to. — Taylor Lorenz

> `17:15` Well, my buddy Claude doesn't think that it's funny.

> `17:15` Don't cancel me because people get so mad when you talk about it. — Taylor Lorenz

> `20:00` this guy's an idiot

> `20:00` open your refrigerator, open up your Google Auth and put in your six-digit code, and then I'm back looking at the screen. — Joe Weisenthal

> `22:45` monetizing confusion — Rachel Karten

> `25:09` Twitter is like a wonderful fairyland compared to Threads. — Taylor Lorenz

## Mentioned Adam Mosseri, Anthropic Claude, Asana, Bloomberg, Bloomberg Screentime, CAA Vault, Canva, Carmen Rodriguez, ChatGPT, Chipotle, Condé Nast, Creative Artists Agency (CAA), Dash Bennett, Dupixent, Federal Trade Commission (FTC), Fruit Love Island, Gilley Cooks, Google, IDA Ireland, Instagram, Instagram Reels, Jaxson Dart, Joe Weisenthal, Kel Brooks, Kevin Lozano, LinkedIn, Link in Bio, Meta, Microsoft Copilot, Minky Couture, Morris Cafritz, Nike, Odd Lots, PayPal Open, Rachel Karten, Salesforce, Schneider Electric, Taylor Lorenz, The Vermont Hollywood, Threads, TikTok, Twitter / X, User Mag, Vanity Fair, YouTube

## Video description Social media is really strange these days: Increasingly, our feeds are starting to look uncanny, as AI slop proliferates all over the Internet. And for the professionals who are tasked with making content, whether it is social media managers or creators, it has been a head-spinning time of change. In this conversation, recorded at the Vermont in Hollywood, we get a primer on social media's AI era with two veteran Internet commentators, Rachel Karten (social media consultant and the author of the Link in Bio newsletter) and Taylor Lorenz (founder of User Mag). We discuss why people are afraid to admit that they love some genres of slop, why the AI discourse is evolving so rapidly, and whether our time online is moving away from scrolling feeds to engaging with chatbots. Read more: <a href="https://www.bloomberg.com/news/articles/2026-09-23/us-rebukes-australia-for-plan-to-target-social-media-algorithms?utm_medium=referral&utm_source=podcast&utm_campaign=odd_lots&utm_content=article" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">US Rebukes Australia for Plan to Target Social Media Algorithms</a> <a href="https://www.bloomberg.com/news/articles/2026-09-21/meta-s-new-muse-ai-app-tops-charts-draws-strong-early-reviews?utm_medium=referral&utm_source=podcast&utm_campaign=odd_lots&utm_content=article" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">Meta’s New Muse AI App Tops Charts, Draws Strong Reviews</a> Only <a href="http://Bloomberg.com" data-inline-card="" data-card-data="">http://Bloomberg.com</a> subscribers can get the Odd Lots newsletter in their inbox each week, plus unlimited access to the site and app. Subscribe at  <a href="https://www.bloomberg.com/subscriptions/oddlots?in_source=oddlotspodcast" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">bloomberg.com/subscriptions/oddlots</a> <a href="http://bloomberg.com/subscriptions/oddlots" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">Subscribe to the Odd Lots Newsletter</a> Join the conversation: <a href="https://discord.gg/oddlots" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">discord.gg/oddlots</a>See <a href="https://omnystudio.com/listener">omnystudio.com/listener</a> for privacy information.

[Source](https://podtrac.com/pts/redirect.mp3/tracking.swap.fm/track/UVBrz8bN8aM2Xe47PEPu/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/8a94442e-5a74-4fa2-8b8d-ae27003a8d6b/7f51b2eb-0524-4cf5-b654-b4ce0132f318/audio.mp3?in_playlist=982f5071-765c-403d-969d-ae27003a8d83)

中文简报

营销人员为应付老板审批伪造AI提示词，社交信息流充斥AI垃圾随着企业管理者将创意判断外包给聊天机器人，合成垃圾内容挤占公开信息流，社交媒体正逐渐放弃个人分享转向广播式娱乐，而真实的情感联结则退守至私信之中。《Odd Lots》主持人 Tracy Alloway 与 Joe Weisenthal 邀请互联网文化专家 Rachel Karten 和 Taylor Lorenz，共同分析生成式人工智能如何颠覆创作者经济与企业营销。讨论详细探讨了“AI脑”管理模式对员工士气的打击——高管在未经培训或缺乏指引的情况下将创意审美外包给聊天机器人，以及受众的双重标准：人们对网络疯传的AI“水果垃圾”视频颇为宽容，却惩罚投机取巧的商业品牌。他们评估了代理机构的实际应用案例与文案创作失败教训，反驳了关于AI存在性风险的末日叙事，并探讨了向泛在计算的转变。最终，这场讨论揭示了社交信息流如何分化为由非粉丝触达驱动的算法娱乐广播，促使真正的人际连接退入私信和群聊中。高管对生成式AI的采纳在企业营销中引发了普遍的“AI脑”现象，管理者在缺乏正规培训或指导规范的情况下将创意判断推给聊天机器人，严重挫伤了创意团队的士气。 (5:00) 大语言模型无门槛的答疑解惑能力正导致教学和实用型内容创作者迅速面临被淘汰的风险，而提供准社会交往体验和独特视角的创作者仍保有一道防御护城河。 (10:00) 垃圾账号正利用平台的发现算法每日发布数十条AI生成的短片，目前多个浏览量极高的 Instagram Reels 均依赖合成画面，以此通过观众的困惑来变现。 (22:45) 公开社交媒体信息流已完全演变为针对非粉丝发现而优化的广播式娱乐，而真实的人际社交互动已几乎全部转移到私信之中。 (30:00)

Reading summary

On Bloomberg's Odd Lots podcast, internet culture analysts Taylor Lorenz and Rachel Karten argue that generative AI is destabilizing digital marketing by encouraging erratic corporate mandates and synthetic spam, while shifting creator survival toward serialized personality and driving genuine social interaction into private channels. Managerial AI Mandates and Creative Disempowerment (5:00) In corporate workplaces, artificial intelligence is primarily arriving as an unguided mandate from executive leadership rather than an organic productivity booster. Drawing on interviews with roughly 100 marketers for her Link in Bio newsletter, Rachel Karten reports that supervisors are offloading their creative instincts onto large language models—including one manager who submitted an employee's meme to Anthropic's Claude to ask if it was funny, and another worker who began faking a Microsoft Copilot interface just to secure sign-off on original ideas. Unlike past enterprise software rollouts that included months of structured training, marketing teams face ad hoc demands that employees describe as demoralizing. Karten and Taylor Lorenz observe that while generative tools offer utility for visual translation—such as assembling quick pitch-deck mockups to illustrate abstract campaign concepts—using language models for copywriting yields bland copy. Furthermore, Karten argues that creative craft serves as an operational proxy for consumers; if a consumer brand automates its public-facing creative output, audiences frequently assume the business is taking similar cost-cutting shortcuts with its physical ingredients and core products. The Double Standard of Synthetic Content (22:45) Audiences display an inconsistent paradox when evaluating AI-generated material. Lorenz notes that viewers routinely enjoy low-stakes, automated humor—such as AI-generated 'fruit slop' soap operas or surreal collision videos—without evaluating the underlying craft, yet they express outrage when a corporate brand uses synthetic imagery to cut human corners. Hostility typically surfaces not from the technology itself, but when viewers realize they have been deceived. Platforms now face an influx of automated spam accounts publishing up to 50 times per day to exploit algorithmic distribution. Karten highlights that several entries among the top 20 most-viewed Instagram Reels of all time are fully synthetic clips designed to mimic raw behind-the-scenes movie footage. By exploiting visual markers of real-world authenticity, these creators engineer high view counts by monetizing audience confusion, which panelists warn compromises long-term platform credibility. Utility Content Vulnerability Versus Serialized Personality (10:00) The rapid advance of conversational AI is splitting the creator economy along lines of utility and personal perspective. Informational creators who built businesses around instructional, 'how-to' guidance—such as basic home repairs, software troubleshooting, or early YouTube-style beauty tutorials—are uniquely vulnerable to obsolescence as conversational assistants like ChatGPT become frictionless diagnostic reference manuals. Conversely, leverage is consolidating around creators offering non-replicable commentary, strong opinions, and serialized episodic formats. Lorenz and Karten emphasize that generic lifestyle broadcasting is obsolete. Successful creators now structure channels like episodic television programs, exemplified by a restaurant line cook (Gilley Cooks) who gained over 500,000 followers in a few months by documenting kitchen shifts through smart glasses. Concurrently, traditional talent agencies like the Creative Artists Agency are building digital archives, such as the 'CAA Vault,' to store verified voice and likeness models for future synthetic licensing. Algorithmic Broadcast Feeds and the Retreat to DMs (30:00) Major social platforms have evolved from personal community networks into passive, broadcast entertainment channels. On Instagram, algorithmic recommendation engines prioritize short-form Reels and Carousels to reach non-followers, sidelining standard single-image posts to reach only legacy followers. If media lacks broad entertainment value, platform recommendation engines fail to distribute it across public surfaces. Citing observations from Instagram head Adam Mosseri at Bloomberg Screentime, the panel underscores that interpersonal social connection has largely vacated the open feed. Rather than broadcasting personal life updates or family photographs publicly, users now reserve authentic sociality for Stories and private Direct Messages, where they share memes and external videos directly with close friends. The discussion underscores an unresolved tension in digital media: while distribution algorithms reward synthetic, high-volume broadcast entertainment, the resulting feed degradation and managerial overreach alienate creative labor and erode user trust, accelerating the migration of authentic human connection away from public feeds and into closed, private channels.

精炼阅读摘要

彭博社旗下播客《Odd Lots》在好莱坞录制的现场节目中指出，生成式人工智能的泛滥与平台算法机制的变迁，正对企业社交媒体运营者与内容创作者的职业生态构成深层冲击。科技记者Taylor Lorenz与行业顾问Rachel Karten认为，管理层对AI的盲目推崇正挫伤员工士气，而自动化低质内容的大规模扩散也迫使创作者转向连载化、人格化和私域化的生存策略。企业内部的“AI脑”困境与创意代理权丧失 (5:00) Rachel Karten在其针对约100名营销从业者的调研中发现，企业管理层在缺乏培训、规范与共识的情况下强推生成式AI工具，引发了从业人员所称的“极度沮丧”情绪。部分管理者甚至将Anthropic的Claude作为判断创意是否幽默的裁判，导致员工不得不使用虚拟的Microsoft Copilot界面包装自身方案以获取审批。Karten指出，以往企业引入如Asana等协同工具时通常会投入数月制定规范与工作流程，而当下AI的部署则高度无序。此外，Karten强调品牌创意历来是衡量企业运营严谨度的外在指标，若连锁快餐或大型消费品牌依赖合成图像做营销，消费者往往会顺理成章质疑其物理实体产品的用料与质检水准。受众对AI内容的双重标准与创作者的类型分化 (10:00) 受众在对待生成式内容时表现出显著的矛盾心态：一方面，用户对无足轻重的人工智能搞笑梗图（如AI生成的蔬菜真人秀《Fruit Love Island》）保持极高包容度甚至频繁转发，但若品牌在商业物料中使用合成内容，受众便会予以强烈谴责。Lorenz补充说明，消费者并非排斥所有AI内容，而是在产生被欺骗感时才会爆发负面情绪。在创作者生存层面，AI的冲击因内容形态而异：纯粹提供信息和维修教程的“工具型创作者”（如水暖维修或基础化妆教学）极易被ChatGPT等大型语言模型替代；相比之下，依托强烈个人观点、独特生活视角以及能够建立准社交互动的个人化创作者，其护城河依然稳固。实操边界、虚假炒作与公共讨论的焦点转移 (17:15) 在日常工作实践中，从业者并未将大语言模型普遍用于文案撰写或脚本策划，而是将其限制在极为特定的辅助环节。Karten指出，AI在创意机构中最有效的用途是在提案方案中生成视觉示意图，帮助缺乏想象力的客户具象化抽象概念；Lorenz则仅将AI用于SEO元数据生成、图片画布扩展等机械性任务，坚决不用其起草正文以保留核心表达。谈及围绕AI的技术争议，Lorenz批评社会舆论过快将底层代码人格化，混淆了法律问责边界与政策制定。针对所谓的“人类灭绝风险”（X-risk），她认为其缺乏现实逻辑，并指出当下反AI力量严重分裂：关注数据中心物理用地与电网负荷的环境批评派，与痴迷于超级智能毁灭人类的风险派之间存在本质目标错位，始终未能形成合力。 “混乱变现”、Threads环境恶化与算法机制重构 (22:45) 平台生态正在被自动化生成的内容农场侵蚀。Karten披露，部分自动化账号为钻算法空子，单日发布多达50条内容以博取爆款几率；在Instagram播放量历史排名前20的Reels短视频中，已有两至三条完全由AI合成，此类视频往往通过伪造“电影幕后实拍花絮”等真实视觉线索制造认知混乱，依靠观众的困惑来收割流量。在Meta的新产品Threads上，由于推荐算法频繁将用户引入脱离语境的攻击性回复，加上AI新闻摘要频发幻觉并捏造品牌争议，该平台被Karten与Lorenz评为当前最具毒性的社交网络。面对主流信息流对无聊内容的算法边缘化，创作者必须像制作连载电视剧一样输出主题鲜明的情节，并利用Instagram优先推荐Reels与轮播图给非粉丝的算法特征去吸纳新受众。信息流向纯娱乐演化与社交关系的私域转移 (30:00) 公开信息流的属性已从熟人社交彻底转变为纯粹的被动广播娱乐。Instagram负责人Adam Mosseri此前在行业会议上指明，社交平台的“社交”本质早已不再是通过公开信息流分享家庭动态或个人生活，而是全面转向私信（DM）及快拍（Stories）。用户当前的典型社交行为是消费公共信息流中的搞笑短视频或梗图，并通过私信向好友定向转发生发互动。在硬件形态层面，关于传统智能手机屏幕使用时间的道德恐慌可能被环境计算（如智能眼镜、耳机及可穿戴硬件）所稀释，但这种演变也可能在削弱传统“算法刷屏”的同时，催生出人际孤立与新型交互摩擦。生成式AI的普及与算法机制的演化正彻底重构社交媒体的生产与消费逻辑：公开信息流已被算法驱动的泛娱乐与AI低质内容占据，迫使创作者向栏目化内容转型，而真实的人际互动则退缩至私信等私域渠道。在此背景下，企业盲目使用AI替代人类审美的管理偏差，以及平台通过制造认知混乱来维持流量的短期行为，正持续消耗商业品牌的信誉与用户对数字生态的基础信任。

Detailed notes — zh

## Background 本期节目由彭博社（Bloomberg）旗下知名播客《Odd Lots》于 2026 年 9 月 24 日在各大音频平台发布。节目由主持人 Tracy Alloway 和 Joe Weisenthal 在好莱坞 The Vermont 剧场现场录制，围绕社交媒体进入“AI 废料”泛滥时代后，邀请两位资深互联网文化观察家 Rachel Karten 与 Taylor Lorenz 共同探讨职业创作者与社媒运营团队的生存震荡、受众对 AI 内容的矛盾心态，以及在线互动从公共信息流向私信与聊天机器人迁移的深刻转变。 ## TL;DR 对话深入拆解了生成式 AI 浪潮下创作者经济与营销职场的剧烈震荡。一线运营人员正遭遇管理层盲目推崇 AI、“把幽默感外包给算法”的荒诞职场困境，而大众在道德上指责品牌偷工减料的同时，又对荒诞离奇的“AI 快餐内容”欲罢不能。随着基础教程类内容被 AI 工具迅速替代，单纯炫耀日常的生活方式网红正在退场，唯有极具真实人设、强情感联结与“连续剧化”叙事的内容才能建立壁垒。此外，传统平台的公共信息流已彻底沦为纯算法驱动的娱乐广场，真正的社交关系全面退守至私密群聊与私信互发梗图之中。 ## Atmosphere & style 作为一场在好莱坞剧场录制的线下 Live 现场秀，整场交流节奏明快、笑声不断，充满了主持人与嘉宾之间默契机敏的抛梗与互动。讨论在剖析“AI 脑老板”、可乐薄荷糖爆炸奇观等怪象时充满了辛辣讽刺与黑色幽默，将一线从业者的心酸与荒谬感展现得淋漓尽致。随着话题切入平台信任危机与人际交互的异化，现场基调在啼笑皆非的戏谑中逐渐沉淀出对技术狂潮下职业命运与真实人性纽带的严肃反思。 ## Bulletin - `05:07` 生成式 AI 正以自上而下的方式深度重塑社交媒体营销工作流，部分中高层管理人员甚至出现将审美与幽默感判断完全外包给 AI 的“AI 脑”异化现象，一线员工甚至被迫将真实创意包装成 AI 界面风格以求通过审批。 - `06:40` 受众对 AI 内容呈现出分裂态度：消费者对品牌使用廉价 AI 素材削减成本持强烈抵触，但对“水果爱之岛”等纯娱乐性质的低质脑洞梗（AI slop）表现出极高的消费黏性与宽容度。 - `10:00` 传统“如何做（How-to）”类实用技能创作者正面临来自 AI 搜索与助手的全方位替代，而依托真实生活经历、独特观点与强拟社会关系（Parasocial）的人格化博主壁垒反而更显稳固。 - `26:15` 社交媒体推荐机制与分发逻辑发生根本转变：静态生活方式炫耀贴已失去流量推荐，算法如今只奖励具有连贯剧集感的微综艺内容，而真实社交互动已全面从公域退守至私信（DM）与限时动态（Stories）。 ## Key takeaways - `00:00` 当内容生产边际成本趋近于零时，职业社交媒体运营团队与网红面临深刻的定位重构与生存危机。 - `02:40` 社交媒体目前的整体内容氛围被称为“糟糕透顶”，充斥着荒诞奇观类的低质 AI 生成视频（AI Slop）。 - `05:07` 管理层的“AI 脑”导致一线职场异化，出现主管用 Claude 判定幽默梗图、员工反向伪造 Copilot 生成界面的荒诞审批流。 - `06:40` 品牌如果在公域物料中使用粗制滥造的 AI 图片，会严重透支消费者对底层产品质量与诚信度的信任。 - `10:00` 大众对 AI 内容存在双重标准：反感商业品牌的偷工减料，却在算法流中对荒诞无厘头的 AI 肥皂剧快餐欲罢不能。 - `10:00` 修马桶、洗碗机等传统实用教程创作者面临 AI 助手的降维打击，而强情感共鸣的拟社会人际交互内容依然具有不可替代性。 - `15:00` AI 舆论周期轮转极快，地方数据中心能耗争议与超级智能灭绝人类的存在性风险阵营诉求割裂，未形成统一政治共识。 - `17:30` AI 文案在实际营销中极难产生打动人心的灵性表达，营销从业者主要将其用于视觉提案预演与繁琐的机械格式支持。 - `20:00` 相较于充满尖酸批评与人身攻击的传统社媒评论区，聊天机器人的无攻击性促使部分用户的屏幕时间发生分流。 - `22:45` AI 自动化账号以每日发布多达 50 条的频率进行“拉老虎机”式算法试错，高播放 Reels 中已出现混淆真实影视幕后的 AI 伪造内容。 - `22:45` Meta 旗下的 Threads 因算法强推愤怒回复区与幻觉频发的新闻摘要，被批评为高度催化网络群殴的极化平台。 - `26:15` 单纯炫耀日常的生活方式博主时代已死，能够像连续剧一样每天连贯更播的第一人称微剧集内容成为吸粉新密码。 - `26:15` CAA 等顶尖经纪公司成立专门项目为旗下明星建立数字克隆体（digital clone），推动名人人格的授权化与批量生产。 - `26:15` Instagram 算法将 Reels 短视频与轮播图完全作为触达非粉丝的推荐燃料，无法吸引陌生公域受众的内容将迅速沉寂。 - `30:00` 现代社交媒体的“社交”定义已全面异化：公域主流彻底沦为纯娱乐广场，真实人际联结退缩至私信互发 Meme 梗图。 ## Detailed walkthrough ### `00:00` 赞助资讯与好莱坞现场：AI冲击下的创作者经济在节目正式开场前，几则赞助商广告展现了当下技术工具与商业环境的变迁：PayPal 聚焦于正在经历考验的“智能体商业”（Agentic Commerce），强调其凭借 25 年结账经验、全球 4 亿个人账户及反欺诈保护为跨端交易提供信任；ChatGPT Work 则主打跨应用和文件的自动化执行能力，旨在将混乱的原始资料转化为可供审阅的初稿；万事达卡（MasterCard）则强调为中小企业提供网络安全威胁防御工具。随后，主持人 Tracy Alloway 与 Joe Weisenthal 宣布，本期《Odd Lots》是在洛杉矶好莱坞的 The Vermont 剧场录制的第一场线下现场秀。谈到洛杉矶，绕不开的话题便是网红经济与创作者经济——在这个城市里，随处可见拥有 500 万甚至 800 万 Instagram 粉丝、但主流大众却闻所未闻的创作者。当下创作者们正在面临一个核心的生存命题：当人工智能让内容生产变得近乎免费、门槛降至极低时，那些以此为生的人究竟该如何应对？ - **录制地点**：洛杉矶好莱坞 The Vermont 演出场馆（Odd Lots 首场洛杉矶现场录制）。 - **核心议题**：当内容生产边际成本趋近于零时，职业内容创作者与社交媒体运营团队的生存处境。 --- ### `02:40` 嘉宾登场与当下社交媒体的“窒息氛围” 为了拆解这一现象，节目邀请了两位深耕互联网文化与社交媒体生态的资深观察家：独立科技媒体 *User Mag* 的创始人 Taylor Lorenz，以及专注社交媒体行业观察、撰写知名 Substack 通讯 *Link in Bio* 的顾问 Rachel Karten。Rachel 介绍，自己的通讯如今已成为各大品牌社交媒体团队的核心行业读物，她经常深度访谈一线运营人员，挖掘各类网络爆梗背后的操盘细节。Taylor Lorenz 则长期追踪科技政策、平台机制以及光怪陆离的在线流行文化。当被问及当前社交媒体的整体氛围（Vibes）究竟如何时，两位嘉宾给出的答案干脆而沉重：“糟糕透顶”（Terrible）。为了说明当下网络内容的荒诞与光怪陆离，Taylor 甚至分享了两人在后台候场时看到的一条 AI 生成视频：一辆可口可乐卡车发生爆炸，喷涌出漫天可乐，有人拿着水桶去接；而紧挨着它的竟然是一辆曼妥思薄荷糖卡车，一个人失足掉进薄荷糖堆里，瞬间被冲天喷发的汽水柱炸向外太空。该视频配文调侃道：“这就是你妈妈会转发给你的那种 AI 视频”。主持人 Joe 笑着打趣，现在的年轻网友或许已经忘记，在古早互联网时代，网民们连续好几年盯着屏幕看的唯一内容，其实也就是可口可乐加曼妥思的喷发实验。 - **嘉宾阵容**： - **Rachel Karten**：社交媒体顾问，行业通讯 *Link in Bio* 主理人。 - **Taylor Lorenz**：科技记者，Substack 通讯 *User Mag* 创始人。 - **后台荒诞 AI 视频**：可口可乐卡车遇上曼妥思卡车的 AI 脑洞爆炸狂欢，折射出当下算法信息流中泛滥的低质奇观内容（AI Slop）。 --- ### `05:07` “你的老板得了AI脑吗？”：荒诞职场与被外包的直觉玩笑过后，讨论切入了专业社交媒体营销团队面临的严酷职场现实。Joe 询问一线运营者到底是将 AI 视作抢饭碗的威胁，还是能成倍放大工作效率的得力工具。Rachel 指出，对于绝大多数营销人员而言，AI 的侵入通常并非源自个人自发探索，而是“自上而下”由管理层强行灌输的。Rachel 曾就此撰写过一篇题为《你的老板得了“AI脑”吗？》（*Does Your Boss Have AI Brain?*）的热门文章，调研了上百位营销一线员工。调查发现，许多中高层管理人员不仅自己深度沉迷于生成式 AI，更开始将最宝贵的人类直觉与审美判断全面外包给模型。 ```mermaid flowchart TD accTitle: 荒诞的 AI 审批流对比 accDescr: 展示一线创作者在正常模式与应对 AI 脑老板时的畸形工作流 subgraph NormalFlow["常规创作流程"] A["策划创意 / 制作 Meme 梗图"] --> B["提交管理层审批"] B --> C["管理层基于经验与直觉评估通过"] end subgraph AIBrainFlow["“AI脑”异化流程"] D["创作者构思真实创意"] --> E["利用 Copilot UI 伪造生成界面"] E --> F["提交给迷信 AI 的老板"] F --> G["老板误以为是 AI 产出予以批准"] D -.-> H["若直接提交原案"] H -.-> I["老板喂给 Claude 提问: '这好笑吗？'"] I -.-> J["算法消解幽默，原案被毙"] end ``` 这种现象引发了极其荒诞且令人窒息的工作场景： 1. **算法判定幽默**：有社交媒体经理花大量心血制作了一张紧跟时事的幽默 Meme 梗图发给主管审核，主管竟然直接将图片扔进 Claude，敲入提示词询问：“请问这个好笑吗？”现场主持人和观众直呼“这实在太凄惨、太令人作呕了”（That's so grim）。 2. **逆向“伪装”成 AI**：另有一位营销人员透露，由于老板只迷信并采纳由微软 Copilot 提供的方案，他不得不先用自己的大脑构思好优质创意，再手工做图把内容伪造成 Copilot 界面的生成样式，提交给老板后才能顺利过审。受访员工在采访中给出的评价是“令人极度心灰意冷”（Demoralizing as hell）。Rachel 强调，这并不意味着营销团队应该彻底封杀 AI，核心症结在于管理层毫无章法的粗暴推广。回想她早年在康泰纳仕（Condé Nast）工作时，公司为了推行项目管理工具 Asana，耗费了数月时间建立标准化培训、理顺工作流规范；而在如今的 AI 浪潮下，没有任何操作哲学或防护栏，一线创作者只能任由被“AI 脑”支配的老板折磨。 --- ### `06:40` 品牌捷径与“水果爱情岛”：大众对AI内容的双标与分裂面对职场上的 AI 乱象，主持人 Tracy 提出了一个反向假设：大众当下对 AI 生成内容表现出普遍反感，但也许再过三到五年——甚至只需短短三个月——大众心理就会发生彻底逆转，变得完全不在乎内容是否由 AI 制作。既然如此，我们现在为什么非得排斥 AI 生成的营销内容？ Rachel 从品牌信任度与消费者预期的角度进行了解读。消费者对一个品牌的认知与其展现出的创意投入深度绑定；品牌在社交媒体创意上倾注的心血，往往被大众视作其产品品质的一面镜子。像耐克（Nike）这样的巨头，其品牌护城河很大程度上就建立在顶级创意的持续投入上。相反，如果消费者看到自己常去的连锁快餐品牌（比如 Chipotle）为了省钱开始使用粗制滥造的 AI 生成配图，内心难免会产生严重的信任动摇：如果他们在面向公众的创意物料上都疯狂走捷径，那他们在看不见的后厨食材和制作流程中，又偷工减料了多少？ ``` 消费者心理信任链路： [ 精品创意 / 人工投入 ] ==> [ 品牌重视体验 ] ==> [ 产品质量值得信赖 ] [ 粗糙 AI 配图 / 走捷径 ] ==> [ 品牌缩减成本 ] ==> [ 质疑产品生产偷工减料 ] ``` 然而，Taylor Lorenz 提出了另一个看似矛盾的残酷事实：在内容消费端，受众其实非常喜欢看 AI 内容，甚至常常乐在其中。例如在 YouTube 上，如今大量博主直接照本宣科念诵由 AI 生成的脚本，评论区里成千上万的观众却赞不绝口，纷纷留言夸奖“深度解析太精彩了”、“文案太强了”。受众反感的往往不是 AI 内容本身，而是“被欺骗的感觉”。这也直接引出了当下网络信息流中极其精神分裂的奇观：用户可能前一秒还在某家品牌的社交动态下义愤填膺地敲下抗议留言，指责品牌方使用 AI 作图“令人作呕、毫无灵魂”；但紧接着手指往下一滑，下一条视频赫然是用 AI 制作的“水果爱情岛”（Fruit Love Island）短剧——讲述一颗多汁的橙子与一根胡萝卜陷入爱河的肥皂剧狗血故事——这群用户却看得津津有味、欲罢不能。大众一方面在道德层面抵制 AI，另一方面又在算法信息流中为各种荒诞的“AI 垃圾快餐”（AI Slop）疯狂贡献点击与时长。 ### `10:00` 梗图的双重标准与创作者的生存分化在探讨大众对 AI 垃圾内容（AI slop）的复杂心态时，嘉宾指出用户往往存在显著的“双标”现象：像“水果版《爱之岛》”（Fruit Love Island）这类荒诞无厘头的 AI 梗图，大家只觉得是低成本的娱乐消遣，根本不会苛求其中的工艺水平，甚至乐呵呵地随手转发给朋友；可一旦一家品牌或商业公司把 AI 生成的平庸内容搬上菜单或广告，消费者就会立刻出离愤怒，痛批其毫无诚意与工匠精神。这种心理机制揭示了受众对商业实体与纯娱乐梗图在“真实性”和“付出感”上的预期差异截然不同。面对“在 AI 泛滥的世界里，拥有现有粉丝基底和真实人设的头部博主会变得更有价值，还是会被新一轮零门槛创作浪潮淹没”这一提问，嘉宾认为这完全取决于内容的本质属性： ```mermaid flowchart TD accTitle: 社交内容类型在AI冲击下的价值分化 accDescr: 展示工具教程类内容与拟社会人际交互内容在面对AI冲击时的不同受力路径 subgraph Disrupted ["工具与教程类 (高可替代性)"] A["基础技能教学生态 (如修洗碗机/修马桶/基础化妆)"] --> B["AI 搜索与多模态助手 (更权威、直观、即时响应)"] B --> C["传统How-to创作者面临受众流失"] end subgraph Resilient ["拟社会关系与独创性 (高壁垒)"] D["深度个人化内容 (生活碎片、真实观点、强烈风格)"] --> E["不可轻易复刻的真实人设与洞察"] E --> F["受众粘性上升与信任溢价"] end ``` - **工具教程类（How-to）内容的衰落**：早年 YouTube 依靠美妆教学、家电修理等实用教程普及推动了媒介平权，但如今面对“如何修洗碗机”这类具体问题，AI 能够以更高的权威性和更快的速度提供解决方案。嘉宾笑着提到自己的母亲前不久还很自豪地用 ChatGPT 自行修好了马桶，这类内容创作者在未来将面临巨大的生存挤压。 - **拟社会关系（Parasocial）与深度表达的壁垒**：那些围绕个人生活经验、独特见解与深度陪伴感建立起的情感纽带，因其极难被算法复刻而显得越发珍贵；优秀的创作者若能利用 AI 工具提升产能，反而能巩固自身的不可替代性。 --- ### `15:00` 舆论周期轮转：数据中心争议与存在性风险迷思互联网舆论对 AI 话题的关注点以惊人的速度迭代流转。上个月人们还在围绕 AI 数据中心的能耗与土地争论不休，随后话题又迅速跳跃到前沿模型本身的技术争议上，公众对科技议题的注意力窗口极为短暂。针对“关注数据中心是否已经过时”的调侃，Taylor Lorenz 分析道： - **选民议程的错位预测**：有政治顾问曾指出，许多人误以为数据中心及算力基建带来的环境与资源问题会是 2028 年大选的议题，实际上它在 2026 年的中期选举周期中就已成为焦点，而到了 2028 年公众大概率又会被新的危机吸引。 - **割裂的反 AI 阵营**：担忧数据中心对本地资源破坏的现实派，与担忧超级人工智能毁灭人类的“存在性风险派”（X-Risk），两者的诉求与立场南辕北辙，至今未能形成任何连贯的政治联盟。谈及广为流传的“AI 终结人类”论调，Taylor 直言：“我们大概率不会有事，即便想彻底消灭地球上的所有人类，在现实层面上也是极其困难的。”她强调更应审视的是人们对 AI 的过度拟人化（Anthropomorphization）倾向：把主观能动性归咎于纯粹的计算机程序，混淆了法律责任归属的讨论边界。正如现场引用的幽默反驳所言，不能因为某物能杀人就认定它更聪明——“病毒杀死了很多人，并不代表它比人类聪明”。 --- ### `17:30` 营销从业者的真实实践：告别文案生成，回归视觉预演在实际工作流中，所谓“让 Claude 判断创意幽默与否”的设想常常令人啼笑皆非。对于营销团队与创作者究竟如何务实地应用 AI，Rachel Karten 与 Taylor Lorenz 达成了高度共识： | 应用场景 | 实际落地情况与态度 | |:--- |:--- | | **文案与发帖点子生成** | **基本不推荐**。业内极少有优秀的社交媒体经理能靠 AI 生成有灵性的好文案，机械化的表述缺乏人情味。 | | **视觉概念预演（Pitch Decks）** | **高度有效**。向缺乏想象力的上级或客户汇报线下落地大创意时，用 AI 快速渲染效果图，极大地降低了沟通阻力。 | | **琐碎辅助与运营支持** | **日常主力**。用于 SEO 关键词优化、在 Canva 等工具中快速进行图像扩图、处理繁杂格式等机械性辅助工作。 | Taylor 特别打趣提醒大家“讨论在工作中使用 AI 时可千万别把我取消（Cancel）了”，因为当下网络舆论对此极为敏感。但她和 Rachel 都强调，在文字创作上自己始终坚持亲手书写以表达真实思想，AI 最大的效用仅仅是帮人越过繁重杂乱的准备工作，充当一个让脑海中创意可视化的辅助手脚。 ### `20:00` 聊天机器人 vs 社交刷屏：屏幕时间的演变与环境化交互对话首先围绕一个核心假设展开：用户曾经花在社交媒体上无休止“刷屏”（scrolling）的时间，是否正在被与聊天机器人（chatbot）的对话所蚕食，从而对依赖传统流量的品牌与营销人员构成威胁？现场嘉宾对此持有不同看法。一种观点认为，AI 对屏幕时间的侵蚀并不是坏事，真正的威胁反而是人们开始渴望彻底“下线”（log off）、远离手机和技术。然而，社交媒体与聊天机器人的功能本质完全不同：社交媒体本质上是娱乐，是给真实的朋友发私信（DM）；而与聊天机器人交谈则是另一种完全不同的功用。主持人打趣反驳道，在 X（前 Twitter）上发帖往往会招来一大堆网民谩骂，比如直接被骂“这家伙是个白痴”或“Joe 已经疯了”；但如果去用 Claude 或 ChatGPT，AI 绝不会用这种刻薄的方式羞辱用户。既然能避开在 Instagram 评论区被人公开嘲讽的痛苦，很多人自然会觉得和聊天机器人对话更令人身心愉悦。嘉宾们指出，Twitter/X 的生态机制与 TikTok、Instagram 大相径庭，后者对大多数人来说更像不需要互动的“电视”，不会面临同等的攻击压力。围绕社会上对“屏幕时间”的道德恐慌（moral panic），讨论进一步延伸到硬件形态的终结：未来大家可能不再整天盯着一块方形屏幕，而是通过各种穿戴设备（如智能耳机、智能眼镜、甚至最新推出的可以对着轻声低语的手环）以及智能家居实现“环境化交互”（ambient interaction）。尽管这一趋势是否真的能减少屏幕依赖仍存争议——正如主持人打趣道：“搞不好以后打开家里的智能冰箱，你还得先掏出手机看 Google Authenticator 输六位双重验证码（2FA）。” ```mermaid flowchart TD accTitle: 用户交互界面的演变路径 accDescr: 展示用户从传统社交媒体屏幕互动转向聊天机器人与环境化硬件的逻辑 subgraph S1["传统社交平台 (2010s)"] A["手机屏幕刷流 (TikTok / Instagram)"] B["公开社交互动 / 评论区 (易遭遇恶意攻击与批评)"] end subgraph S2["新兴交互模式"] C["AI 聊天对话 (Claude / ChatGPT 友好且不嘲讽)"] D["环境化硬件 / 可穿戴设备 (智能眼镜 / 智能家居)"] end A -.->|"注意力分流"| C B -.->|"避开社交攻击"| C A -.->|"形态转变"| D ``` - **社交平台功能分化**：TikTok 与 Instagram 演变为纯娱乐性质的“电视流”，而 Twitter/X 仍是强观点碰撞与人身攻击的高发地。 - **对话界面的吸引力**：相较于社交媒体上尖酸刻薄的网民，AI 助手提供了无攻击性、更具确定性的交互反馈。 - **环境交互替代论**：随着语音穿戴设备（如近期发布的可低语手环）和智能家电普及，部分观点认为用户“低头看屏幕”的物理动作正在被环境感知所稀释。 --- ### `22:45` AI 废料时代的平台困局与 Threads 的“毒性”爆发面对海量涌现且几乎零成本的“AI 废料”（AI slop），平台与用户的关系正站在十字路口：一方面，被劣质 AI 内容劝退的用户可能选择逃离至 Discord、私信小群等封闭圈子；另一方面，平台算法的分发与策展（curation）能力被迫卷到极致，必须精准识别用户到底喜不喜欢这些合成内容。嘉宾直言，老牌社交平台动作迟缓、应对乏力，特别是 Meta 依然深受 2010 年代传统社交媒体逻辑的束缚，地位犹如一头影响力见顶的“恐龙”。更令人担忧的是，AI 账号正以机器的极速与规模优势大肆生成内容——有些账号一天能发 50 条帖子，就像在 Instagram 算法的老虎机前疯狂拉下拉杆，只为赌中一次爆款。在近期 Instagram 历史播放量前 20 的 Reels 中，竟有两到三条完全由 AI 生成。其中一条视频虚构了电影拍摄现场的巨浪模型摧毁房屋的幕后花絮（BTS），借由“揭秘真实幕后”的外衣行彻底的虚假欺骗之实。这种靠混淆现实来“变现困惑”（monetize confusion）的手段，短期内虽能收割流量，长期看却在极具破坏性地透支平台的信任基石。随后，话题转到了 Meta 旗下的文本平台 Threads。尽管它常向非活跃用户推送通知以刷存在感，但在嘉宾的犀利点评中，Threads 被直接冠以“当前最毒社交媒体”（the most toxic social media）的称号。其推荐机制倾向于直接把用户丢进极其愤怒的评论回复区，极速催化群体极化和网络围殴（pile-on），其毒性甚至让 Twitter/X 看起来都像一片“美好的仙境”。雪上加霜的是，Threads 内部署的 AI 自动新闻摘要经常满嘴跑火车（hallucination），把荒唐的品牌八卦当成事实煞有介事地总结展示，形成了一幅怪诞混乱的内容图景。 - **AI 账号的工业化轰炸**：部分账号凭借单日发布 50 次的频率，通过“拉老虎机”式的高频测试算法阈值赚取曝光。 - **爆款 Reels 的 AI 渗透**：历史播放榜前列已出现 AI 生成的伪造影视“拍摄幕后”（BTS），反映出以“伪造真实感”牟利的内容套路正在蔓延。 - **Threads 的推荐机制争议**：依靠把用户推向高情绪对立与愤怒争吵来维系活跃度，加上漏洞百出的 AI 新闻摘要，使其生态环境急剧恶化。 --- ### `26:15` 2026 年的创作者生存法则：告别生活方式，拥抱“连续剧化” 进入 2026 年，对于那些迫切希望涨粉的普通创作者而言，旧有的社媒玩法已经彻底失效。两位嘉宾给出了极为鲜明的分水岭判断：单纯“展示生活”的生活方式网红（lifestyle influencer）时代已经终结，大众对平淡炫耀日常的生活贴不仅毫无兴趣，甚至充斥着强烈的负面情绪。如今真正能迅速吸粉的内容，核心在于极具连贯性与熟悉感的“连续剧化”（serialized content）——创作者必须把自己的账号经营成一档每天准时播出的微型“电视节目”。嘉宾举了一个典型例子：一位餐厅厨师佩戴 Meta 智能眼镜，用第一人称视角记录自己每天的烹饪过程（如账号“Gil Holy cooks”），在短短两三个月内就收获了超过 50 万粉丝。观众每天守着他的内容，就像追看一部固定的日常剧集。与个人创作者的剧集化转型平行演进的，是名人与顶级机构对数字分身的制度化接纳。好莱坞顶级经纪公司 CAA 已经设立了名为“CAA 秘库”（CAA Vaults）的项目，专门收录旗下明星的数据，打造其官方数字克隆体（digital clone），以便在声画授权下批量生成内容。这也引发了主持人的调侃：未来彭博社或许也能无限生成自动播客版的《Odd Lots》。在具体的运营策略上，根据 Instagram 官方透露的推荐算法逻辑： - **Reels 短视频与 Carousels 轮播图**：目前被系统专门用于触达“非粉丝”（non-followers）受众。 - **推荐引擎的核心燃料**：当前平台算法最看重的是能否引发非粉丝的观看与互动；如果一条内容只有原本的存量关注者才能看懂，系统将迅速停止向公域流量推荐。 ### `30:00` 社交互动的私密化转型与节目尾声讨论聚焦于当下各大社交平台内容分发的残酷现实：如果一段内容不够具有娱乐性，平台算法便根本不会将其推送到主信息流（Main feed）的公共推荐池中。单张静态图片（single image posts）的受众面日益收窄，基本沦为只有既有关注者才会浏览的存量内容；而平台真正的“社群氛围与人际联结”，则已全面退守至限时动态（Stories）与私信（DMs）这类更私密的空间。嘉宾特别提及此前在彭博 Screen Time 大会上，Instagram 负责人亚当·莫塞里（Adam Mosseri）发表过的一段发人深省的言论：如今社交媒体当中的“社交”（social），其定义早已不再是公开发布一张家庭生活照，而是“在私信中向朋友转发一件能让你联想到对方的东西”。这种转变被评价为“既令人感到遗憾，却又无比清晰地揭示了当下在社交媒体上‘社交’的真正含义”——维系人际纽带的载体不再是彼此人生中的重大近况更新，而变成了私信框里来回发送的一张张搞笑梗图（Memes）。在好莱坞 The Vermont 剧场录制的本期《Odd Lots Live》现场对话至此落下帷幕。主持人特蕾西·阿洛威（Tracy Alloway）与乔·威森塔尔（Joe Weisenthal）向深入拆解了 AI 浪潮如何冲击内容创作者与社交媒体运营者的两位嘉宾——《User Mag》创始人泰勒·洛伦兹（Taylor Lorenz）以及《Link in Bio》作者兼社交媒体顾问雷切尔·卡滕（Rachel Karten）致以谢意，盛赞这一领域的演变极其引人入胜。随后，节目公布了主创团队的社交媒体账号、每日新闻通讯以及 Discord 交流社区，并播送了包括美式橄榄球四分卫贾克森·达特（Jaxson Dart）联名款 Minky Couture 毛毯，以及即将于伦敦探讨 AI 规模、安全与技术主权的 Bloomberg Tech 峰会赞助信息。 - **社交互动载体的迁移**：公共信息流演变为纯娱乐化广场，真实社交与社群互动深度转移至 Stories 和私信（DMs）。 - **亚当·莫塞里（Adam Mosseri）的核心论点**：现代社交媒体的“社交”行为已演化为私信互发 Meme 梗图，而非公开分享个人现实生活。 - **嘉宾与主持信息**：特蕾西·阿洛威（@tracyalloway）、乔·威森塔尔（@thestalwart）、泰勒·洛伦兹（@taylorlorenz）、雷切尔·卡滕（@milkkarten）。 - **制作团队致谢**：卡门·罗德里格斯（Carmen Rodriguez）、达希尔·贝内特（Dashiell Bennett）、凯伦·布鲁克斯（Kaelen Brooks）与凯文·洛扎诺（Kevin Lozano）。 ## Notable quotes > `02:40` 这就是你妈妈会转发给你的那种 AI 视频 — Taylor Lorenz > `05:07` 请问这个好笑吗？ — Rachel Karten > `05:07` 令人极度心灰意冷 — Rachel Karten > `15:00` 我们大概率不会有事，即便想彻底消灭地球上的所有人类，在现实层面上也是极其困难的。 — Taylor Lorenz > `15:00` 病毒杀死了很多人，并不代表它比人类聪明 — Taylor Lorenz > `17:30` 讨论在工作中使用 AI 时可千万别把我取消（Cancel）了 — Taylor Lorenz > `20:00` 这家伙是个白痴 — Joe Weisenthal > `22:45` 当前最毒社交媒体 — Taylor Lorenz > `30:00` 在私信中向朋友转发一件能让你联想到对方的东西 — Adam Mosseri ## Mentioned Bloomberg, Odd Lots, Tracy Alloway, Joe Weisenthal, Rachel Karten, Taylor Lorenz, The Vermont, PayPal, ChatGPT Work, MasterCard, Instagram, User Mag, Link in Bio, Substack, Claude, Microsoft Copilot, Condé Nast, Asana, Nike, Chipotle, YouTube, 爱之岛 (Love Island), Canva, X (Twitter), TikTok, Google Authenticator, Meta, Threads, Reels, Discord, CAA, CAA Vaults, Adam Mosseri, Jaxson Dart, Minky Couture, Carmen Rodriguez, Dashiell Bennett, Kaelen Brooks, Kevin Lozano ## Video description Social media is really strange these days: Increasingly, our feeds are starting to look uncanny, as AI slop proliferates all over the Internet. And for the professionals who are tasked with making content, whether it is social media managers or creators, it has been a head-spinning time of change. In this conversation, recorded at the Vermont in Hollywood, we get a primer on social media's AI era with two veteran Internet commentators, Rachel Karten (social media consultant and the author of the Link in Bio newsletter) and Taylor Lorenz (founder of User Mag). We discuss why people are afraid to admit that they love some genres of slop, why the AI discourse is evolving so rapidly, and whether our time online is moving away from scrolling feeds to engaging with chatbots. Read more: <a href="https://www.bloomberg.com/news/articles/2026-09-23/us-rebukes-australia-for-plan-to-target-social-media-algorithms?utm_medium=referral&utm_source=podcast&utm_campaign=odd_lots&utm_content=article" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">US Rebukes Australia for Plan to Target Social Media Algorithms</a> <a href="https://www.bloomberg.com/news/articles/2026-09-21/meta-s-new-muse-ai-app-tops-charts-draws-strong-early-reviews?utm_medium=referral&utm_source=podcast&utm_campaign=odd_lots&utm_content=article" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">Meta’s New Muse AI App Tops Charts, Draws Strong Reviews</a> Only <a href="http://Bloomberg.com" data-inline-card="" data-card-data="">http://Bloomberg.com</a> subscribers can get the Odd Lots newsletter in their inbox each week, plus unlimited access to the site and app. Subscribe at  <a href="https://www.bloomberg.com/subscriptions/oddlots?in_source=oddlotspodcast" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">bloomberg.com/subscriptions/oddlots</a> <a href="http://bloomberg.com/subscriptions/oddlots" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">Subscribe to the Odd Lots Newsletter</a> Join the conversation: <a href="https://discord.gg/oddlots" data-prosemirror-content-type="mark" data-prosemirror-mark-name="link">discord.gg/oddlots</a>See <a href="https://omnystudio.com/listener">omnystudio.com/listener</a> for privacy information. [Source](https://podtrac.com/pts/redirect.mp3/tracking.swap.fm/track/UVBrz8bN8aM2Xe47PEPu/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/8a94442e-5a74-4fa2-8b8d-ae27003a8d6b/7f51b2eb-0524-4cf5-b654-b4ce0132f318/audio.mp3?in_playlist=982f5071-765c-403d-969d-ae27003a8d83)

Source quotes

[2:40] Taylor Lorenz: 这就是你妈妈会转发给你的那种 AI 视频 [5:07] Rachel Karten: 请问这个好笑吗？ [5:07] Rachel Karten: 令人极度心灰意冷 [15:00] Taylor Lorenz: 我们大概率不会有事，即便想彻底消灭地球上的所有人类，在现实层面上也是极其困难的。 [15:00] Taylor Lorenz: 病毒杀死了很多人，并不代表它比人类聪明 [17:30] Taylor Lorenz: 讨论在工作中使用 AI 时可千万别把我取消（Cancel）了 [20:00] Joe Weisenthal: 这家伙是个白痴 [22:45] Taylor Lorenz: 当前最毒社交媒体 [30:00] Adam Mosseri: 在私信中向朋友转发一件能让你联想到对方的东西

Source quotes

[5:00] the AI videos that your mom sends you [5:00] Is this funny? [5:00] demoralizing as hell [15:00] Oh, you're still arguing about data centers? Wasn't that a July issue? That was so July. [15:00] Taylor Lorenz: First of all, it's not going to kill us. It's actually so hard to kill all humans on Earth, even if we wanted to. [17:15] Well, my buddy Claude doesn't think that it's funny. [17:15] Taylor Lorenz: Don't cancel me because people get so mad when you talk about it. [20:00] this guy's an idiot [20:00] Joe Weisenthal: open your refrigerator, open up your Google Auth and put in your six-digit code, and then I'm back looking at the screen. [22:45] Rachel Karten: monetizing confusion [25:09] Taylor Lorenz: Twitter is like a wonderful fairyland compared to Threads.

Original transcript

[0:00] Did you ever notice how you spend hours shopping online, only to pause at checkout because you wonder if you trusted enough to hit buy now? Agentic Commerce is testing that moment more than ever. That's where PayPal comes in. With 25 years of checkouts, 400 million consumer accounts globally, and the benefit of fraud protection. So no matter where a purchase starts, it ends with trust. Built for payments, growth in a Gentic. PayPal Open. Built for all business. Visit paypalopen.com. Some people treat ChatGPT like some kind of smart search engine, and some use it to get work done. ChatGPT Work is a new way of working in ChatGPT that can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work. It's designed to help you move from a chaotic starting point to a reviewable first version. So all the source materials, briefs, and scattered information that you have to grind through to turn into something useful This is Robert Smith from Business History. If you're listening to this, there's a good chance you're a small business owner. And like every small business owner, you started with a dream. to do what you love and watch it grow. What you probably didn't dream about? Keeping up with cyber threats. That's where MasterCard can help with access to tools that help identify cyber threats to better protect your business. Building your dream business. Priceless. For cybersecurity in a changing world, There's MasterCard. Learn more at MasterCard.com slash smallbusiness. Hello and welcome to another episode of the All Thoughts Podcast. I'm Traci Allaway. And I'm Joe Weisenthal. So we're here in Hollywood, our first ever live Odd Thoughts show in LA. And when we come to LA, I feel like we have to talk about the influencer economy, right? The influencer economy, the creator economy, people who have 8 million followers on Instagram that none of us have ever heard of. There's so much I don't know about this world. There's so much I need to learn. And of course, we need to learn how like AI is upending all of it. That's right. So this is the existential question facing creators. What happens to them when creating content essentially becomes like free or extremely low cost? So we have two very special guests for you. We are going to be speaking with Taylor Lorenz. in bio. since we're here in LA and as we've alluded to in the past, We've got to learn about influencers and people with 5 million followers on Instagram and how they're... how they're dealing with this moment. Yeah. So we do, in fact, have the perfect guest. That's right. Thrilled to welcome to the stage the one and only Taylor Lorenz, founder of UserMag, as well as Rachel Carton, social media consultant, author of the great sub-static LinkedIn bio. Hello. Thank you both so much for doing this. What do you start? What do you maybe either one of you? What are you covering these days? How would you describe your coverage? Sure. I write a newsletter called Lincoln Bio. Oh, yeah. It's really good. Everyone should subscribe. Thank you. It's about working in social media. It's the primary trade publication for social media teams. And I interview experts. to Chanty Beluga. I forgot about that one. Is that still... That was like a one week forever. And I am a tech journalist. I write usermag.co. That's my sub stack. I write about I kind of do a lot of roundups about like online culture, write about tech, a lot of tech policy stuff, too. A little bit of everything. So everybody subscribe to that one as well. Absolutely. What are the vibes like right now in social media? It's a loaded question. Terrible. Yeah. All right. Go on. Backstage, Taylor showed me a video that she was impressed by. No, shut up. You guys. She was really impressed by it. Literally, we were backstage, and I haven't been on Twitter that much today. And I was like, [5:00] Did you guys see a Coca-Cola truck exploded? It was a Coca-Cola truck that was exploding Coke out and somebody had a bucket. And she was like, did you see this? I said, Taylor, please. And then the next scene is there's actually a Mentos truck next door. And the guy falls in and like explodes through the sky. And I was like, oh, my God, I'm becoming. And the caption was like the AI videos that your mom sends you. Oh, man. And I showed Rachel. I think younger people might not remember the fact that for many years, the only content on the internet was Mentos and Coca-Cola. That's all we looked at for a few years. That's really strange. No, but here's an important question. So the people that you talk to who are social media professionals, goals. Do they feel as if like AI is we're going to like, how is it about like, are they mostly of the frame? Like, oh, I can do what I do. Or do they mostly see it as an augmenter? Oh, I can do much more of the job that I know how to do for years with these new tools. So I think the way that AI is showing up for a lot of marketing teams is through their boss I wrote an article about this called Does Your Boss Have AI Brain And I talked to like a hundred marketers who are dealing with an issue where their boss is asking them to use AI and their boss is using AI themselves and they're outsourcing all of their gut instincts to AI. And so I heard stories of a social manager who worked hard on a funny meme and they'll send it to their boss and they run it through Claude and ask, is this funny? That's so grim. That's so grim. I heard another example of a marketer who their boss only buys into ideas that came from Copilot. And so they actually will mock up their ideas into the Copilot UI to send their boss to get it approved. And so the quote that sticks with me from that article is demoralizing as hell. And so that's the way that a lot of marketers are interacting with AI, is their boss sort of forcing it upon them. And my takeaway from that wasn't like ban AI from marketing teams, but I felt like the way that AI is being implemented is sort of a free for all. And I was thinking back to my time when I worked at Condé Nast, I think they had Asana onboarded and we spent months right Like learning how to use Asana what the right way to use this workflow And that training isn happening on marketing teams And so it feels like there no philosophies or guardrails for marketers to feel like their boss is you know giving to them Devil's advocate question. I know we're talking about how grim it is to have to work for a boss with AI psychosis and everyone at the moment hates, seems to hate AI generated content, But you could imagine a world in three to five years, maybe even sooner, three months where we don't. Yeah, three months where we don't actually care anymore. Like, why should we dislike AI generated social media or marketing content? Well, I think I can speak to it from the brand side. And I do feel like consumers have an expectation that a brand is putting creative effort into things. and the effort that they put into their social media is a reflection of the effort that they are putting into their product. And I think that consumers feel that through it, and that has always been true about creative. You know, brands like Nike have built so much of their world around their investment into creative and so it seen as a shortcut and it might make you ask you know if uh your favorite chipotle whatever your favorite fast food chain is using ai generated imagery what other shortcuts might they be taking in their process yeah i totally agree i think from a content standpoint too people actually do enjoy ai generated content um it's so crazy on youtube i don't know if you guys have noticed but all these youtubers are just reading ai generated scripts now And it's so and you look at the comments, people like amazing analysis, great script. Like he's so I think people actually like it sometimes when they don't know they're consuming it. Sometimes they actively know they're consuming. I think people just don't want to feel tricked. That's when it starts to feel bad because you're like, oh, yeah, I think it was your newsletter that I first heard about the what is it? The fruit slop or the fruit soap. Yeah. And I think you made the point. you can have a bunch of people in a comment section that's like, oh, I'm so disappointed. Oh, you used AI for this, it's disgusting. And then they scroll to the next reel, and it's like, oh, I love this story about an orange falling in love with a carrot. Fruit Love Island. Yeah, Fruit Love Island, exactly. What do you, explain this disconnect as you see it? Well, I think that consumers are confused in general. [10:00] over what they should like online right now. And I think that, Brute Love Island feels like a low stakes, like it's a meme, and they don't care about the craft of a meme, but they care about the craft of a company that they purchase from. And so I think, yeah, they could be the same person who's yelling at the company for an AI-generated meme. menu and they're sending their friend, AI fruit slot. Yeah. Um, I sort of asked... Hayes this question earlier, but in our AI generated world, I guess. How do we feel about existing content or existing incumbents when it comes to social media? Do people who already have an audience, who already have a following, who maybe have proved authenticity in some way or that they're willing to put effort into their own work, do they become more valuable in this environment or is it another case of like low entry costs, anyone can come in and make content for free. And so we see a sort of further trend of, I guess what we used to call the democratization of media. Well, I think that, like... If you think about that, I think it depends on the type of content you're making. Are you making deeply parasocial content about like your life and your opinions? So many creators, the value that they like, the way that that media was democratized was through how to content. If you think of like even just the earliest YouTube videos, it was all makeup videos, how to do this. And I actually think AI is a better authority and people will turn to AI for. how to fix your dishwasher. Like it's much easier to get that. So I think those creators are going to struggle. But then there are other creators that that are going to use the tools to make great content and they make content that is not easily replicable. You know, maybe it's commentary that's sort of my mom by the way she was very proud she was telling me that she used chat gpt to fix her toilet recently for her which was i wouldn't have expected her to be able to do that Did you ever notice how you spend hours shopping online, only to pause right before you hit buy now? That tiny hesitation, the one where you wonder if it is trustworthy, can make or break the sale. Now, with AI changing the way consumers discover and compare things, that split-second trust question matters more than ever. That's where PayPal comes in. For more than 25 years, PayPal has been a leader in online payments. And now they're at the forefront of agentic commerce, making it work for businesses, letting them maintain control of their brand and their customer relationships. still feels familiar and deeply dependable. Build for payments, growth, and agentic. PayPal open. Built for all business. Visit paypalopen.com to get started. Some people treat ChatGPT like some kind of smart search engine, and some use it to get work done. ChatGPT Work is a new way of working in ChatGPT that can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work. It's designed to help you move from a chaotic starting point to a reviewable first version. So all the source materials, briefs, and scattered information that you have to grind through to turn into something useful Something useful. Put ChatGPT to work on your most ambitious ideas and projects. Get started at ChatGPT.com by selecting Work Mode, available on Plus and Pro plans. Are your ad campaigns lighting up the dashboard, but not the pipeline? That's bullspend, and marketers are calling it out in Dashboard Confessions. My boss asked for results, so I opened my dashboard for the only positive-sounding metric I had. Impressions. See revenue, not just reach. LinkedIn delivers the highest return on ad spend of major ad networks. Advertise on LinkedIn. Spend $250 on your first campaign and get a $250 credit. Go to linkedin.com slash campaign. Terms and conditions apply. Even if you're using prescription creams for your eczema, you still may be hearing this. That relentless itching that makes you want to scratch and scratch and scratch and scratch. When your moderate to severe eczema isn't well controlled, it's time to ask about Dupixent, Dupilumab. Dupixent helps block a key source of inflammation inside the body that can cause eczema, helping adults stay ahead of it with clearer skin and less itch. Allergic reactions, including skin reactions, can occur. Get help right away for face, mouth, tongue, or throat swelling, wheezing, or trouble breathing. Tell your doctor right away of new or worsening eye problems like eye pain or vision changes, skin symptoms, joint aches and pain, or a parasitic infection. Don't change or stop other treatments without talking to your doctor. Dupixent helps you feel the pain. [15:00] See the difference for yourself. Talk to your eczema specialist, visit dupixent.com or call 1-844-DUPIXENT. - Taylor is a keen observer of the online discourse. You know, you're maybe more online than I am somehow. I'm just curious, like, You know, in the last months, You know, a month ago, there was like a lot of people arguing about data centers. And then in the last week, there were a bunch of people arguing about models themselves as this sort of like site of controversy. Do you feel like, let's just say we're talking about AI that like, the speed with which the public takes on a new thing, is it's just like at extraordinary rapid pace. And that's like, oh, you're still arguing about data centers? Wasn't that like a sort of, Wasn't that a July issue? That was so July. I feel like there's... I mean, the social... There's going to be outrage on social media. Data centers is... an issue of 2026. I saw, I think it was like Morris Katz, the political consultant actually saying like, people are going to make a mistake and think it's a 2028 election cycle issue. It's actually a very 2026 election cycle issue and probably by 2028, we will have something new to get mad about. But I think, yeah, the data center discourse has been interesting because those people are against AI for a very different reason than the X-Risk people that think it's going to kill all of us. And so far, those people have not really come together to form any sort of coherent coalition. Yeah. So I feel like maybe they should join up, but they have very different interests. And I don't think it's going to kill us, by the way. We're going to be fine. I was going to ask, so you wrote about the idea of when we talk about rogue AI potentially killing us, getting the discourse and the language about that possibility, right? Give us the sort of nutshell argument. Like, what is the distinction that we need to be making here? Okay, so first of all, it's not going to kill us. It's just like, it's actually so hard to kill all humans on Earth, like, even if we wanted to. So... So we're probably going to be fine. I think like what I think is concerning or worth interrogating is the way that we anthropomorphize AI. And that's what I wrote about is the way that people assign agency to these products. And I think in some ways, it's really great to do that because we have to describe AI You know, you have to sort of describe how they work, but they're not humans. And we are talking about computer programs. And I do think that when we go to these policy debates about liability and stuff, it's really important to recognize that these are not. like truly living beings. - I saw one of the former FTC commissioners. He's like, AI isn't gonna kill us. We survived the Black Death and that only killed 50% of us. So it's like, all right, great. That's a really big-- - But also, just 'cause something kills us doesn't mean it's smarter than us. You know, like a virus can kill a lot of people. It's not smarter than us. - I'm sure that's comforting. - Okay, on the question of intelligence. So like, yes, this idea, like it makes you sick to your stomach to think like, oh, you've come up with something funny. And then they're like... Well, my buddy Claude doesn't think that it's funny. He's just like, oh my God. Do you see in the marketing world, currently productive uses of the technology, whether it's A/B testing things or like, okay, like it's easy to make fun of it, but also, AI models do often tend to like do pretty like, good answers to things. Do you see positive, productive applications of it in the marketing world? The, the... Interesting way that I've seen it is that a lot of agencies will be selling through ideas that they want to do in the real world and in their decks or in a way to help a boss who has maybe a lack of imagination visualize what that big idea is. Okay. I have been interested by using AI to mock something up to then be able to make it in the real world and have like a... human impact on it. But so that's that's something that I'm interested in. But I don't I don't hear that a lot of marketers are having success having it generate captions or post ideas. And it's not really a strategy that I personally recommend either. Taylor, how are you using AI? Good question. Don't cancel me because people get so mad when you talk about it. But I use it mostly, kind of like what Rachel was saying, as like an assistive tool. I don't think it's a good writing. Also, if I'm writing, I kind of want to express my thoughts. Like, that's the whole point. So I mostly use it for like SEO captions, generating, like, you know, making on the Canva, you know, app, making an image bigger, things like that. podcast. So we're, you know, there's, there's a gap, you know, Yeah. One thing I'm curious about for both of you from the perspective of social media is – We have 24 hours in the day. Let's say eight of them were sleeping roughly. They're 16. So it's like, OK, maybe you can scroll for 16 hours. [20:00] They got to do a few other things. But now there's like a bunch of that time that once it was spent scrolling might be spend time talking to a chatbot. And I'm curious if you feel it where it's like, Setting aside whether AI is useful, not bad in creation, whether chatbot time is taking away from scrolling time to the detriment of brands and marketers. No. I think getting off of phones and off of technology is more of a threat. But AI is going to get us off the phones. Like, I do think screen time is going to drown. Because a lot of people are chatting using chatbots, right? Social media is entertainment. Social media is DMing your friend something who's a real person. And I think chatting with the chatbot is a very different function. And I'm more concerned about people wanting to log off. - But when I go to social media, I tweet something, and a bunch of people are like, "Oh, this guy's an idiot. "People told me that today." Or Joe, I used to like Joe, but he's lost his mind. - Well, maybe you're using AI too much. - I probably am, but what I'm saying is, I could then go, if I use Claude, it's never gonna insult me like that. So it's like, maybe I'll just-- - Well, that's bad. That's bad. - No, of course it's bad, but I have the iron to resist the temptation. But couldn't a lot of people say, wow, this is much more pleasant. I don't get dunked on when I talk to... ChatGPT in the way I do... in the comments section on Instagram? I would say that Twitter as a social media platform is so significantly different than TikTok and Instagram where people are scrolling and it's like TV for them and they're not getting insulted on there. And so I think that's a little bit more specific to a Twitter. But don't you think that like screen time, I think it's, I write a lot about the moral panic over screen time and everybody is so concerned about it and it is going away. Like, I do think that like this form factor of this type of phone is going to die quite soon. you know, people with the ear pods in, like, I just think we're spending less and less time actually looking at a screen. And I think in the future, we're going to interact much more ambiently. I think a small subset of people want us to believe that personally. I don't see that out in the world. I guess I see that so much where I just see, I see like, I get what you're saying for entertainment. But in terms of just like, like your mom, like going to fix the dishwasher, like just all these other times where we used to have to turn to a screen to text it, to message a friend. But that's a screen though. Right. But I'm saying like, if you're asking about like screen, like looking at our screens, I just think that like, we're looking at our screens less and less. And if you check it less, you're going to get sucked in less. to learn how to fix a toilet is actually, I would have called a plumber and had a real interaction with somebody. I'm just saying, like, if you're arguing a screen... I think that I just think there's less and less reasons to be on our phone as we have wearables, as I'm sure you guys saw. Like there was this persona band that came out this week that you can like whisper to. I feel I just feel like everyone's fridge and washing everything in your home is smart now. Like, I just feel like we're not going to have to do everything through our phone as much, maybe with the glasses, too. And so I just think that that might end up altering the way that we interact with these media platforms. antisocial in my opinion. Oh, yeah. No, totally. Yeah, totally. I'm not saying that it's more pro-social. We're all agreed on the antisocialness. Don't worry. We all agree it's bad. You're going to need dual-factor authentication to open your refrigerator. Like, open your refrigerator, open up your Google Auth and put in your six-digit code, and then I'm back looking at the screen. Wait, can I just ask, on the platforms themselves, in the age of AI slop and essentially free content And I can kind of, I could see arguing it from both ways. So like, On the one hand, Maybe a bunch of people are going to want to migrate away from X or Instagram if it's just full of AI and they're not really into that. Maybe we see more people going to DMs and chat groups and discords and check out the All Thoughts Discord, by the way, if you aren't already on it. But on the other hand, maybe curation becomes even more important and maybe having an algorithm that like can really tell what you like and your preference for AI slop or not becomes even more important. How are you thinking about that? You can go. Um... I thought you were going to answer first, so I was like, I'm going to have time to think. I think, well, so I do think that the social media platforms are not, they're like on their back foot, right? Like, they're not, they're not. They're not moving fast enough. I do think like in some ways, meta, I mean, Google's better positioned than meta. Meta is in a really hard place right now because I think they are very like 2010s sort of like social media. But I think, I don't know, it'll be, it'll depend on how people start to view entertainment and how, like how much good content can be on that app. But it's kind of a dinosaur. I don't see meta like having the same level of influences. I think it kind of peaked in the 2010s. Hmm. I mean, I think in the short term, what AI is doing to [25:00] - Platforms is a little bit concerning, just in the sense that they can there's like sort of a scale and quickness that AI accounts. I'll go to look at an AI account that can post, you know, 50 times a day. They're just like pulling the slot machine of Instagram and they're going to get a hit. And so – I think that... when I looked up recently, what were the top 20 most viewed reels ever? And I think two to three were AI generated, like pretty recent ones. One of them was showing a set and a wave and how it would, they're filming a wave for a movie and how it would impact or houses. And so it was using the framework of like proof of reality or BTS to essentially confuse people of how they film something. I think that's, really bad to uh, essentially monetize confusion. And I think that's a short term benefit with really bad long term results. That's why I think people are going to retreat towards these more... these chat interfaces that are simpler and not so overwhelming and will be more kind of interactive and make them feel less lonely, even though I agree with you. It's not good. Wait, I have a question. Speaking of meta, and I may have asked you this before, but I can never remember. Who uses threads? Who is like, no, I'm like aware sometimes. Like, what is the deal? I still get thread notifications every once in a while. And I'm like, oh, what is this? And then I remember. Why are they like you? Do you have a grand theory? My mom. And why does that appeal to her? And I recently published a hot takes newsletter. And I said that threads is the most toxic social media. Oh, it is by far. and so you'll get sort of dropped off into somebody's really angry reply to something, and that's the what fuels it, essentially. And it turns into a very... quick sort of group think pile on in a way that even worse than Twitter. Joe, it's not Twitter is like a wonderful fairyland compared to threads. Are you serious? It's insane. It also does the AI generated. We were literally just talking about this, but like it makes it threads does the AI generated news summaries. But it's just lying. It just makes things like she was showing me a crazy brand drama on there. But like. It's a crazy place. We only have a few minutes left, so I'm going to ask a question that we sort of alluded to at the beginning, but I'm sure there are a lot of people out in the audience who would like to grow their social media accounts and their presence. But, like, what are your top tips for... in the age of AI for getting to, as Joe says, an account that mysteriously has 5 million followers. What are you... I would say the accounts that I see that are growing right now, all their content has a real like through line of familiarity. It's like they're creating their own TV show. There's a cook at a restaurant who wears meta glasses and just records what he's cooking. I believe he's... grown to like 500,000 followers in maybe two or three months, like Gil Holy cooks. And so that's all that they post is just like, my show is on. They are, making an episode every day. Yeah, it seems to be a lot of like serialized content. It's all becoming shows. You have to really think about entertaining. I think the era of like sharing your life is like very declining. Nobody cares. And I think like, so don't try to be a lifestyle influencer in 2026. It's not there's a lot of anger towards those people. I'm curious though, speaking of the lifestyle influencers, Are people signing deals with like, yes, you can use my voice. You can create a digital clone of me. You can have 10 versions of me. Like, is that happening? Yeah, I wrote a piece for Vanity Fair recently about this. And CIA has like really leaned into this. They have what's called the CIA vaults where like all of their big stars, like, you know, they use AI to like basically cut down on people. And I do think that that is probably the way that that it's going to go. I guess Bloomberg owns the rights to our voices, right? They're going to start generating just endless odd lots. They might do. You know, one thing I like reading about, Rachel, in your sub-sec, it's like, you talk about the various tactics, right? carousels or what are some of the other ones that are what's hot these days in terms of like presentation approaches. Well, so Instagram shared with my audience a really helpful chart that showed what each sort of... format is for. And so carousels and reels are for reaching a new audience. Okay. Those will go out to non followers. And what fuels the algorithm right now is non follower views. If it's something that only your audience understands, they're like, [30:00] That's not entertaining enough. We're not going to put it on the main channel. And then single image posts just are for your audience. And then stories and DMs they see as where community happens. And there was a conversation, I think at Bloomberg Screen Time, actually, where Adam Osseri said that the social and social media is not through sharing, you know, an image of your family anymore, necessarily. It is through DMing your friends, something that reminds you of them. And I thought that was very unfortunate, but clarifying in terms of what it means to be social on social media anymore. through DMs and sending them a meme that reminds them of them more than... an important update in their life. Taylor and Rachel, thank you so much for joining us at OddLots Live. Closing out the show. It's such a fascinating area. Really appreciate talking to both of you. Thank you. You can follow me @tracyalloway. And I'm Joe Waisenthal. You can follow me @thestalwart. Follow our guests. Taylor Lorenz. She's at Taylor Lorenz. And Rachel Carton at Milk Carton. Follow our producers. Carmen Rodriguez at Carmen Armand. Dash O'Bennett at Dashpot. Kale Brooks at Kale Brooks. And Kevin Lozano at Kevin Lloyd Lozano. And for more Odd Lots content, you should check out our daily newsletter. You can find that at Bloomberg.com forward slash Odd Lots. And you can chat about all these topics 24-7 in our Discord, discord.gg slash Odd Lots. And if you enjoyed this conversation, then please leave a comment or like the video, or better yet, subscribe. Thanks for watching and listening. Thank you. The real recovery starts after the final whistle, and that's where Minky Couture comes in. From leading Ole Miss to the pros, Jackson Dart knows peak performance means knowing how to come down, too. That's why he teamed up with Minky Couture to create a collection as tough as his game and as soft as his downtime. When the crowd is gone and your body is still buzzing, Minky Couture is the blanket built for the moments after the game takes everything out of you. Shop the Jackson Dart Collection now at minkeecouture.com. That's M-I-N-K-Y-C-O-U-T-U-R-E dot com. Minky Couture. AI is entering its most consequential phase where scale, safety and sovereignty will determine who leads and who lags. Join Bloomberg Tech in London on November 2nd and 3rd as global leaders across business, finance and policy examine the defining tradeoffs shaping the future of AI. Thank you to our presenting sponsor, Salesforce and supporting sponsors, IDA Ireland and Schneider Electric. Learn more at BloombergLive.com slash Tech London.