TAWS

Thursday, November 12, 2026

Jensen Huang Resigns From Nvidia, United Bans Outside AI Agents, Anthropic IPO Oversubscribed | TAWS

Watch on YouTube · Listen on Spotify

The rundown

AI-generated from the transcript

Jensen Huang resigned as Nvidia CEO, with finance chief Colette Kress stepping in on an interim basis five days before earnings. Austin and Ron also dig into United Airlines banning outside AI agents from booking tickets, Russian pranksters getting three OpenAI national security officials on a video call, and Bloomberg's report that Anthropic's IPO is about six times oversubscribed, partly by its own chip suppliers. The show closes with Workday's argument that employers, not its software, decide who gets hired, and a South Korean tax exam score that set off AI glasses cheating claims.

Jensen Huang resigns as Nvidia CEO

Watch from 2:55

What happened
Jensen Huang resigned as Nvidia CEO and gave up his board seat. Nvidia filed the news with regulators at 7:52 a.m. Central. CFO Colette Kress, finance chief since 2013, takes over on an interim basis while keeping her finance job. The board named its first ever chair, Harvey Jones, and will search inside and outside the company for a permanent CEO. On Sunday the board had expressed confidence in Huang; the Wall Street Journal then reported directors were preparing to fire him. Nvidia stock closed down about 4%. Earnings are still set for Tuesday, November 17. Semafor reported Apple hardware chief Johny Srouji was being considered.
The claim
Nvidia's board says customer deals and the product roadmap will continue without interruption.
Austin's take
Austin bets Kress stays at least through earnings and that the permanent pick will be an insider either way. He argues Nvidia's strong, long-tenured culture should be kept intact, pointing to Apple's insider succession after Steve Jobs. Ron argued for an outsider like Srouji as a clean break.
Why it matters
Nvidia supplies the chips much of the AI industry runs on, and Wall Street gets its first earnings report without Huang only five days into Kress's tenure.
What to do
Watch the November 17 earnings call for signs of how stable the roadmap and leadership really are, and follow who the board names as permanent CEO.
Not confirmed yet
The Wall Street Journal report that directors were preparing to fire Huang and Semafor's report on Srouji are media reports, not confirmed by Nvidia in the episode.

United Airlines bans outside AI agents from booking tickets

Watch from 7:39

What happened
United's new contract bans outside AI agents from booking or changing tickets and lets United cancel such bookings. Travel site Skipped says it's the first US airline to say so openly. Travel blog View from the Wing reports that even the passenger's own approval doesn't exempt the booking, and notes an agent could buy through Expedia, where United would only see an Expedia booking. AI can still show United flights and prices for a human to buy. United CEO Scott Kirby said at a Chicago event, "I don't want to create another middleman."
The claim
United says sharing your login with an outside AI assistant can expose sensitive information and that assistants could make unwanted purchases or spend frequent flyer miles. It also says AI check-in could create regulatory risk because a human must confirm they aren't carrying banned items.
Austin's take
Austin calls it a foul. He agrees a human should confirm the banned items check, but says canceling a purchase the passenger approved goes too far, and questions how United could enforce the rule consistently given the travel site gap. He reads Kirby's comment as United protecting its own sales, similar to other companies like Amazon that resist shopping agents.
Why it matters
If airlines can cancel tickets your AI agent bought with your approval, the promise of agents that shop and book for you hits a wall, and other companies may follow.
What to do
If you use an AI assistant for travel, have it find options but complete United purchases yourself, and avoid handing it your airline login.
Not confirmed yet
It's unclear how United will enforce the rule, and whether other airlines will follow suit.

Russian pranksters fool three OpenAI national security officials

Watch from 8:34

What happened
The Intercept revealed that prank duo Vovan and Lexus posed as Ukrainian official Rustem Umerov and spent seventeen minutes on video with three OpenAI officials. Sasha Baker, OpenAI's head of national security policy, said OpenAI had been in contact with the Pentagon about aiding Ukraine; asked about Elon Musk, she said, "I'm afraid we probably can't help you with Elon Musk." The team discussed work with Ukraine using a security model called Daybreak Blue. Ben Crampton, who leads OpenAI's European national security work, described reluctance among senior staff about offensive military uses. The duo has pranked officials since 2011 and is accused of working with the Russian government, which they deny. OpenAI has a $200 million contract with the US Department of War, though reporting doesn't tie it to this call.
The claim
OpenAI says it opened an investigation weeks ago and concluded the call was a prank. Spokesperson Nate Evans says the company took steps to prevent repeats.
Austin's take
Austin says worry. A known act got seventeen minutes of shop talk with a national security team, and "took steps" needs specifics. His bigger concern is what this means for everyday people if trained officials can be fooled. Ron said laugh, calling it a blooper reel.
Why it matters
Impersonation is getting easier, and if an AI company's security team can be fooled on video, ordinary people and organizations are even more exposed.
What to do
Verify unexpected high-stakes calls through a separate known channel before sharing anything sensitive, no matter how convincing the face or voice.
Not confirmed yet
Nobody has reported an AI-faked voice on OpenAI's call; the AI-mimicked voice claim came from a separate Lithuanian adviser case. The pranksters' alleged ties to the Russian government are disputed.

Anthropic IPO reportedly oversubscribed, with chip suppliers buying in

Watch from 18:03

What happened
Bloomberg, citing unnamed insiders, reports orders cover Anthropic's IPO about six times over, with trading under ticker ANT expected November 19. Anthropic's updated filing lists firms agreeing to buy $26 billion of stock at the IPO price, led by Nvidia and Google at about $10 billion each. AMD is buying $2.5 billion, counting toward a $5 billion commitment tied to a chip deal. Google's check uses an option from a March financing deal. June's private round valued Anthropic at $965 billion; the IPO range puts it near $1.4 trillion. The filing shows $33.2 billion in revenue from January through September, up from $2.5 billion a year earlier, and an operating loss that shrank to $1.3 billion from $5.2 billion. Polymarket has 90% odds the IPO happens by November 19.
Austin's take
Austin calls it too circular: chipmakers sell Anthropic chips and then buy its stock. He sees the business as real and fast growing, but the money underneath as tangled. In a play-money bet he took "no" on an IPO by November 19, while Ron took "yes." Not financial advice.
Why it matters
Suppliers investing in their own customers can make demand look stronger than it is, which matters for anyone judging the AI boom's valuations.
What to do
If you're considering the stock, look at how much of the demand comes from deal-linked partners and how big the jump from the June valuation is.
Not confirmed yet
The six-times oversubscription comes from unnamed insiders via Bloomberg. The November 19 trading date is expected, not guaranteed.

Workday tells court its customers make the hiring calls

Watch from 22:11

What happened
In a new federal court filing in a bias suit brought by Derek Mobley, a Black job seeker over 40 who says he was rejected from more than 100 jobs, Workday says its customers decide who gets hired. Mobley alleges Workday's screening software was behind the rejections. His side wants four nationwide groups to sue together: Black applicants, women, people 40 and older, and people with disabilities. Workday wants all four turned down. An age discrimination collective won early approval in May 2025, and more than 15,000 people joined it. No court has ruled on bias. Workday says it processed more than 356 million applications in 2024. A hearing on the four groups is set for March 9, 2027.
The claim
Workday says its customers, not Workday, decide who gets hired.
Austin's take
Austin says Workday should answer alongside employers, comparing its defense to a help desk transferring the call. If one tool sits in front of that many applications, its maker should answer for how it screens. Ron argued employers who set up and trust the tool are responsible.
Why it matters
This is one of the first big cases targeting the software maker rather than the employer, and it could shape who is accountable when automated screening rejects people unfairly.
What to do
If you suspect you were screened out by automated hiring tools, keep records of applications and rejections, and watch the March 2027 hearing.
Not confirmed yet
No court has ruled on whether bias occurred, and the application total doesn't show how many, if any, were unfairly screened out.

Top South Korean tax exam scorer faces AI glasses cheating claims

Watch from 26:41

What happened
South Korea's top scorer on the tax accountant exam says they missed one of 160 questions, including a perfect tax law score where the average was about 30, with about one minute per question. Online posts claim AI glasses were used, and some accountants called the score abnormal. About 18% passed the first round, down more than four points. HRD Korea, which runs the exam, calls the claims speculation, says no formal complaint was filed, and is reviewing test centers and proctors from the April and July rounds. AI glasses cheating has occurred on separate Korean exams. Seoul's education office trained 239 officers to block smart devices at college entrance exam sites, and the ministry is deploying such officers nationwide for the first time.
The claim
The Tax Accountants Association wants a thorough investigation but warns that judging on score alone risks irreparable harm.
Austin's take
Austin calls checking test sites fair, but says scrutiny of the rooms shouldn't become a public verdict on the person. Exam security needs to catch up with devices, while the review shows restraint.
Why it matters
As AI wearables spread, exceptional results may draw automatic suspicion, putting both exam credibility and innocent high achievers at risk.
What to do
Hold off on judging the candidate until officials finish their review; if you run or take high-stakes exams, expect stricter device checks.
Not confirmed yet
No direct evidence has emerged that this candidate used AI glasses; officials call the claims speculation.

The Loop

The weekly newsletter

The week in AI in one email: what happened, why it matters, what to do about it.