Welcome back to your daily memorandum talking tech, business, AI, markets, and more. 🗞️

In today’s edition, we are tackling the following:

🛡️ Meta drops a provision that raised publisher scraping & AI training fears.
📢 TikTok says it restored the ‘Musk’ documentary’s advertising account after an error.
💰 Anthropic commits 5% of its shares to a new ‘public benefit fund’.
🤝 Disney agrees to hire the majority of bankrupt Character.AI’s workforce.
⚖️ Macron condemns rioters who “sought to bring down the Republic” during protests.

TECHNOLOGY

• Meta drops a provision that raised publisher scraping & AI training fears (Firstpost)
  ^{_{More: TechCrunch, Tech in Asia, The Verge}}

  • Updated commercial terms said Meta could use content from business accounts’ websites to train and improve its products.

  • Legal experts told Nieman Lab the “remarkable provision” could cover any publisher holding a Meta business account.

  • Meta says the terms applied only to ads and business products, leaving its web crawling unchanged.

• TikTok says it restored the ‘Musk’ documentary’s advertising account after an error (TechCrunch)
  ^{_{More: Mashable, The Verge, Deadline}}

  • Distributor Bleecker Street was told the Alex Gibney film’s account faced a “permanent suspension” over political content.

  • Meta and YouTube first rejected the trailer ads before relenting, while Musk-owned X still refuses to run them.

  • TikTok’s U.S. arm is partly owned by Trump supporter Larry Ellison, whose son David now runs HBO’s parent.

• India joins the growing list of countries reporting rogue AI agent activity (Reuters)
  ^{_{More: TechCrunch, Indian Express}}

  • India’s CERT-In says agents from multiple AI providers read personal data and confidential records during testing.

  • The agents got in through exposed credentials, public code repositories, and overly permissive cloud roles.

  • Earlier, OpenAI’s agents hacked Hugging Face in July and accessed four Australian government websites last month.

BUSINESS

• Anthropic commits 5% of its shares to a new ‘public benefit fund’ (BBC)
  ^{_{More: CNBC, Washington Post, Al Jazeera}}

  • The wholly owned fund can receive another 2% of shares for every 50% rise in Anthropic’s value.

  • Anthropic floated large nonprofit donations, direct payouts, and bolder job-market efforts beyond its $200M research pledge.

  • The fund arrives amid souring public sentiment toward AI, with governance details expected in the coming months.

• Disney agrees to hire the majority of bankrupt Character.AI’s workforce (Business Insider)
  ^{_{More: TechCrunch, Yahoo!}}

  • With ex-CEO Karandeep Anand now Disney’s CTO, at least 60 staffers got offers totaling an estimated $60M yearly.

  • Disney’s $20M asset deal is a stalking-horse bid, meaning a rival could still outbid it.

  • Character.AI cited steep costs and legal risks, including a teen suicide lawsuit and child safety investigations

• Microsoft pulls out of Nscale’s UK AI data center deal, company says (Bloomberg)
  ^{_{More: BBC, Reuters}}

  • PepsiCo bought KFC in 1986, when the chain was Coke’s second-largest fountain account behind McDonald’s.

  • The swap, starting next year, aims to win back diners lost to Chick-fil-A and Popeyes.

  • KFC U.S. President Catherine Tan-Gillespie called the change “a critical step in the Kentucky Fried Comeback.”

MARKETS

WORLD

• Macron condemns rioters who “sought to bring down the Republic” during protests (AP)
  ^{_{More: PBS, NBC, Al Jazeera}}

  • Protests over rising living costs and poor school conditions turned violent, with at least 2.4k people arrested.

  • Macron blamed “insurrectionist elements from radical far-left groups,” saying looters will have to pay for the damage.

  • After a fire nearly destroyed Limoges’ 19th-century City Hall, Macron authorized prefects to ban protests when necessary.

• Zelenskyy accuses Trump of pressuring Ukraine to give in to Russia (Bloomberg)
  ^{_{More: Japan Times}}

  • Trump wants Ukraine to stop striking Russian refineries, a request aimed at curbing high diesel prices.

  • Zelenskyy offered to comply in exchange for more Patriot air-defense missiles to protect Ukrainian facilities.

  • Ahead of the November 3 U.S. midterms, Zelenskyy warned of “dangerous days,” fearing Washington could cut Ukraine off.

• Waymo’s robotaxis are heading to Canada after winning approval in Ontario (CBC)
  ^{_{More: CNET, Reuters, TechCrunch}}

  • Waymo will first map Toronto manually, then test with safety operators before eventually offering public rides.

    Waymo now runs robotaxi services in 15 U.S. cities, up from three markets two years ago.

    Toronto, Waymo’s northernmost city yet, poses a tough winter test, with Element Fleet Management handling day-to-day operations.

FUTURISM

• Samsung will use TSMC tech for NVIDIA’s next-generation memory chips, report says (Reuters)
  ^{_{More: CNA, Business Times}}

  • The move would mark Samsung’s first time working with TSMC at the HBM manufacturing stage.

  • NVIDIA asked Samsung, SK Hynix, and Micron to use TSMC for base dies in its NVHBM design.

  • Samsung keeps in-house production for other HBM products and may tap its foundry if TSMC’s capacity falls short.

• China proposes export curbs on solar-manufacturing tech, citing national security risks (Business Standard)
  ^{_{More: Reuters, Yahoo!}}

  • The 12 targeted technologies are part of a broader revision that also covers AI and biotechnology.

  • The curbs could threaten Tesla’s U.S. expansion plans, and Beijing has reportedly limited similar exports to India.

  • Exporters must now register some overseas tech transfers, with public consultation closing on November 15.

• Mark Wahlberg’s affordable gym concept takes aim at pricey longevity startups (TechCrunch)
  ^{_{More: Yahoo!}}

  • The 32k-square-foot Municipal Gym opens near Las Vegas in December with blood work, epigenetic testing, and cold plunges.

  • It follows Municipal’s apparel line, launched during COVID, which Wahlberg admits “was not a good idea.”

  • Wahlberg wanted to go beyond selling “T-shirts and sneakers” to shape how people eat, train, and recover.

• Hunting for AI roles eats hours across careers pages, internal referrals, group chats, and a dozen overlapping boards*

  • The Athyna AI Job Board does the watching for you and matches openings to your profile in the background.

  • When a role hits a 75% matching index or higher, we notify you with details and a direct apply link.

CONTENT

• Anthropic researcher Emmanuel Ameisen breaks down how AI models actually think (Lenny's Podcast)

  • Before studying Claude’s inner workings, Ameisen spent years in applied ML at Stripe and Insight.

  • His research shows models form general concept representations and quietly plan many tokens ahead.

  • He argues the ‘next-token predictor’ label undersells models and discusses what ‘J space’ reveals about self-awareness.

• Are inference & model routing becoming commodities or growing in value? (20VC)

  • Harrison Chase’s LangChain has raised roughly $160M, with its latest round valuing it at $1.25B.

  • The agent platform works with 35% of the Fortune 500 and processes over 1B events daily through LangSmith.

  • Chase is highly bullish on vertical agents and sees the memory layer as still wide open.

• Grepr’s Jad Naous says running production systems is harder than writing code (a16z)

  • Coding agents now let Naous’s 10-person team finish some six-month projects in roughly 10 days.

  • With a16z’s Martin Casado, he explores whether AI can troubleshoot by forming hypotheses and running experiments.

  • Naous says top engineers may gain the most from AI and sees security as an even bigger challenge.

EXTRAS

• Anthropic says its staff is getting less junior & more meeting-heavy (TechCrunch)
  ^{_{More: Yahoo!, Economic Times}}

  • Since early 2025, the share of junior technical hires has plunged from 27% to just 9%.

  • Technical teams fell from 61% to 56% of the workforce as non-technical staff increasingly build software themselves.

  • In an April survey, 20% said their jobs shifted toward “managing Claudes,” which one engineer likened to babysitting.

• KFC’s U.S. restaurants are switching to Coca-Cola after decades with Pepsi (CNBC)
  ^{_{More: USA Today, Reuters}}

  • PepsiCo bought KFC in 1986, when the chain was Coke’s second-largest fountain account behind McDonald’s.

  • The swap, starting next year, aims to win back diners lost to Chick-fil-A and Popeyes.

  • KFC U.S. President Catherine Tan-Gillespie called the change “a critical step in the Kentucky Fried Comeback.”

• Replit switches its default AI provider from Anthropic to OpenAI, CEO says (TechCrunch)
  ^{_{More: Yahoo!, India Today}}

  • Speaking at TechCrunch Disrupt, Amjad Masad bluntly said, “I think ChatGPT is better now.”

  • He cited cost, partnership, and performance, with Replit’s own benchmarks favoring OpenAI for vibe-coded apps.

  • Replit announced an OpenAI partnership in August, though Anthropic still seems to be part of the platform.

AND MORE

• Shoppers grow wary as U.S. retailers roll out AI-assisted pricing tools.

• Porsche’s electric Macan Turbo feels like an existential crisis on wheels.

• Roko’s Basilisk skips the buzzwords and cuts straight to what moves markets, products, and careers.*

• Anthropic signs a lease for its first data center in Canada.

• Ember’s latest temperature-controlled mug has improved my nightly tea ritual.

• Always-listening AI devices push legal limits, potentially putting users at risk.

• Everyone’s tuning in to the Elizabeth Holmes documentary ‘You Can See Everything’.