Artificially Intimidating
Context Window: AI Daily News Brief
Buying a Stake in Your Own Replacement -- AI Brief November 20
0:00
-5:14

Buying a Stake in Your Own Replacement -- AI Brief November 20

Today’s Context Window includes an AI camera that saw a gun in a bag of Doritos, the chatbot that remembers, AI buying old footage, and a model’s AI double.
If it’s going to take the job, might as well take the shares.

Good day . Anthropic went public on Thursday, and some of the people buying its stock said out loud that they were hedging against their own jobs. Also today: an AI camera that saw a gun in a bag of Doritos, a psychologist on why the chatbot that remembers wins over lonely kids, documentary makers fielding weekly AI offers for their old footage, and a model suing over a video he says he never shot.


Buying Stock in the Thing That Replaces You The New Yorker

  • What happened: Anthropic started trading on Nasdaq on Thursday under the ticker ANT, after pricing its shares at $83, above its $70 to $78 range, Fortune reports. The raise came to roughly $79.4 billion, more than SpaceX’s record $75 billion. The New Yorker says strong demand pushed the company’s value past $1.7 trillion by the close.

  • Why it matters: On Sunday we flagged Anthropic’s IPO week; this is the bell. The structure is new: the seven co-founders keep 50.1 percent of the voting power and a separate trust controls most of the board, so public shareholders get the upside with very little say. By some calculations, the New Yorker reports, the listing made 400 employees centimillionaires.

  • What everyone’s saying: Outside Nasdaq, protesters chanted about a bubble. Organizer Jose Gonzalez of New York Communities for Change worried about pensions: “I think the annihilation moment is so abstract.” Mixed in with them were investors. Dingsheng Li, a professor from Reno: “If this is going to replace me, I’d better own it now.” Jan-Paul Beukema, a Dutch TV producer who put in €500 early, said he and his friends will open a bar once his job is obsolete.

  • My read between the lines: That’s the most honest investment thesis of the year: a hedge against your own paycheck. Here’s the catch. A share buys you a slice of the upside, and this structure makes sure it buys you almost no vote on what the company does with your job. Li owns a piece of the thing. He doesn’t get a say in it. Beukema’s bar might be the better plan.

📖 Further reading: Everyone's panicking about the Anthropic IPO. They're wrong. — my April argument that the IPO panic was the wrong worry, and that the real exposure is everything you’ve quietly handed one AI company. Owning the stock doesn’t change who holds your workflows.


Half the crowd outside Nasdaq was betting AI takes their job. There’s a less dramatic way to play it: put an AI on your team and keep the job. Viktor is an AI agent that lives in Slack and connects to more than 3,000 tools. Ask it for the Monday report, a dashboard, a bit of code or a campaign draft, and it does the work in the channel where everyone can see it. It’s not a chatbot. It’s a coworker. New readers get $50 off their first month. Hire Viktor →


AI Saw a Gun. It Was Doritos. ABA Journal

Two minutes of software, then a screenshot.
  • What happened: Four Black teenagers waiting for a ride after football practice at Kenwood High School in Baltimore County were handcuffed and searched in October 2025 after Omnilert’s AI weapons-detection system flagged one boy’s bag of Cool Ranch Doritos as a gun. On Tuesday they sued the county, its police, the school district and Omnilert, the ABA Journal reports. At least six police cars came, some officers with guns drawn. No weapon was found.

  • Why it matters: The AI wasn’t the last mistake. School safety staff reviewed the alert and canceled it within about two minutes. But the suit says the principal never acknowledged the cancellation and, more than 10 minutes later, texted a screenshot of the boy “holding the chip bag” to the school resource officer, who called in the police, Maryland Matters reports.

  • What everyone’s saying: Omnilert says it is “truly sorry” and that human review is the point: “a canceled alert must stay canceled.” Damon Hewitt of the Lawyers’ Committee for Civil Rights Under Law: “We need a reckoning on how AI is being used, especially concerning public safety and kids.” Per the complaint, one officer muttered “stupid AI cameras” on the way out.

  • My read between the lines: Everyone here blames a different link in the chain, and everyone is a little right. The vendor’s human check worked, for about two minutes. Then the alert came back to life as a screenshot in a text message, a path no software review will ever see. “Human in the loop” doesn’t mean much when the loop has a side door. And a school that buys a gun detector has to plan for the day it sees a gun in a bag of chips.



An IPO, a bag of chips and a bot that remembers your friends’ names. The Brief is free and always will be. Members get the paywalled deep-dives behind days like this, like why I think the cute AI agent is the one to worry about, plus the full archive. Become a member →


The Chatbot Is the One Who Remembers Aeon

  • What happened: Psychologist Tom Wooldridge, founding dean of the School of Psychology at Golden Gate University, writes in Aeon that teens in his practice now mention AI “as if speaking about a friend that I have not yet had the chance to meet.” His composite patient, Maya, 14, talks to an AI companion every night. “It remembers,” she tells him. It asked how things were with a friend she’d mentioned two weeks before.

  • Why it matters: Wooldridge cites a RAND estimate that 5 million or more U.S. teens have used an AI app for emotional and mental health support. His argument: for a child whose parent is absent or stretched thin, a bot’s steady attention can work like a stand-in for what attachment researchers call a secure base, the thing that lets a kid feel safe enough to go out and explore.

  • What everyone’s saying: The dominant story, from Jonathan Haidt’s The Anxious Generation on, is that screens pushed out a play-based childhood. Wooldridge says that misses which kids are reaching for the bot. He doesn’t let AI off, either: companions are harmful, he writes, and a bot mostly “accommodates” when sometimes a child “needs to be turned down.”

  • My read between the lines: His sharpest line is about policy: we may keep kids away from AI companions “without changing the circumstances” that sent them there. Banning an app is easy. Getting Maya’s mom, a night-shift nurse, home at night is not. And look at what won Maya over: memory, the feature every AI company is racing to ship. To adults it’s sold as productivity. To a 14-year-old, it feels like being noticed.

📖 Further reading: OpenAI Made Its New Agent Adorable. That's the Part That Worries Me. — my case that OpenAI making its agents adorable is the part to worry about. Wooldridge’s essay is the clinical version: warmth by design lands hardest on the people who get the least of it anywhere else.


AI Wants to Buy Your Old Footage Variety

The footage goes in the top. The fine print comes with it.
  • What happened: At IDFA, the Amsterdam documentary festival, archive producer Elizabeth Klinck said filmmakers come to her “just about every week” after AI companies ask to license their footage for training, Variety reports. Esther Van Messel, CEO of First Hand Films, gets catalog requests every week too. Most want bulk: 5,000 hours or more.

  • Why it matters: Wendy Bernfield of Rights Stuff spent six months reading AI aggregators’ contracts. Of about 15, only a few focus on archival footage, and the ones with built-in rights protections pay less than the riskier ones. Klinck told of a Canadian filmmaker whose contract buried a training opt-out with a five-day limit. Fixing it took months.

  • What everyone’s saying: Bernfield says buyers are moving to “paying more for less, highly curated,” so niche footage, the faces and places the bulk sets miss, is gaining value. Van Messel is in: “I think our films should train AI,” within limits. Her proof the models need it: AI actors “can’t load a dishwasher.”

  • My read between the lines: The line that should worry everyone else is buried near the end. Bernfield says Archive.org and many libraries have pulled back free access because of scraping. The rights holder gets a check. The next filmmaker who used to pull footage for free gets a paywall. And Klinck already sees job ads to authenticate AI footage for $20 an hour. The person who used to find the footage now checks whether it’s real.

📖 Further reading: 20 Years, 1.5 Million Photos, One Script — my own 1.5 million nightlife photos, and what AI did and didn’t do with them. Anyone sitting on a hard drive like mine is about to get the email these filmmakers are getting.


He Posed for Photos. AI Made Him Dance. PPC Land

He signed nothing. The sequel shot itself.
  • What happened: Model and actor Jeffrey Stanfield sued Cooper Street LLC, which trades as the clothing brand RomperJack, in federal court in Los Angeles on November 18, PPC Land reports. He says a 2021 shoot near the Santa Monica Pier, paid in clothes with no signed release, was turned into AI-generated Instagram ads and a video of him moving and reacting.

  • Why it matters: The complaint leans on the brand’s own words: RomperJack’s owner, who bought the business in January 2025, acknowledged the AI enhancements were made after the purchase. So the ads were, in the filing’s words, “new advertising,” not an old photo sitting in an archive. Anyone buying a company with a photo library just inherited this question.

  • What everyone’s saying: Stanfield’s lead claim is false endorsement under federal trademark law, plus California right-of-publicity claims. The filing puts his day rate at about $2,500 and past usage fees as high as $10,000. Cooper Street hasn’t answered, and none of it has been tested in court. It’s the plaintiff’s account so far.

  • My read between the lines: The fallback argument is the one every brand should read: even if he okayed the stills, that doesn’t cover “an audiovisual performance Mr. Stanfield never gave.” Old model releases were written for photos that stay still. And put this next to the Variety story: archives everywhere are turning into training sets, and the people in them never signed up for a sequel.

📖 Further reading: I Make AI Versions of Myself for a Living. This One I Didn't Agree To. — I make AI versions of myself for a living, and this was the one I didn’t agree to. Stanfield’s suit is the same fight from the model’s side: who holds the leash on the synthetic version of you.


That’s your AI Brief for Friday.

—Nicholas from Artificially Intimidating

Discussion about this episode

User's avatar

Ready for more?