Austin covers five stories: OpenAI used its own models to help identify an employee who planted a recorder in its office, Character.AI filed for bankruptcy with Disney opening the bidding at $20 million, ASML says it cannot build enough chipmaking machines through 2028, more than 400 scientists called for an international AI watchdog, and UK graduate employers report record AI cheating in job interviews. A running thread: are we in an AI boom or a bust? Austin's view is that it's still very early.
Watch from 2:51
- What happened
- A routine sweep on September 30 found an encrypted recorder in an OpenAI conference room. Investigators fed the device's storage into OpenAI's models, which pointed to a product employee. The FBI arrested the employee, and the company fired them on October 9.
- The claim
- OpenAI says its models helped investigators identify the person behind the device, and that the arrested employee may have planned to install more bugs.
- Austin's take
- Corporate espionage isn't new, but the stakes around AI raise it to a national security level. He expects more of this across AI labs and wonders how companies will lock down and treat current employees, and whether AI-powered spy catching gets more advanced.
- Why it matters
- The most valuable AI secrets are now targets for old-fashioned spycraft, and labs are starting to turn their own models on security problems.
- What to do
- If you work somewhere with sensitive IP, expect tighter physical security and insider-threat checks. Austin's question for you: could your company spot a spy?
- Not confirmed yet
- OpenAI hasn't said who the spy worked for, and the FBI investigation is still open.
Watch from 7:17
- What happened
- Character.AI, once a billion-dollar chatbot startup, filed for bankruptcy in Delaware, listing assets and debts between $50 million and $100 million. Disney's $20 million opening bid buys a license to some technology, not the whole company. Other bidders can top it before a sale closes around December. Disney sent Character.AI a cease and desist over its characters in 2025.
- The claim
- The company says the app stays online with minimal disruption during the sale.
- Austin's take
- Austin sees it as a warning to 'AI wrapper' startups that put a use case on top of a base model without enough differentiation. In a field this fast, you have to keep shipping something new to compete. He'd like to see the tech find new life at Disney.
- Why it matters
- A former unicorn selling for scraps is a sign that thin apps built on top of frontier models are vulnerable, and it's fueling 'is the bubble bursting' talk.
- What to do
- If you rely on Character.AI, save anything you care about now. If you're building or buying AI products, ask what they offer that the underlying model doesn't.
- Not confirmed yet
- If no buyer wants to keep the app running, it could shut down entirely. Whether this is the first AI bubble casualty is an open question.
Watch from 11:03
- What happened
- ASML, the only company making the machines behind top AI chips, says it can't build enough of them through 2028. Quarterly sales hit a record €11.7 billion, up 55% from last year, and profit more than doubled. It raised its yearly forecast for the third time. Some customers want to reserve machines for 2029 and 2030. Shares rose more than 3% in Amsterdam.
- The claim
- ASML says even a planned 30% boost in output won't cover AI demand.
- Austin's take
- Lithography machines are the wonder machines of the moment. Chip designers like Microsoft, Nvidia and OpenAI send designs mainly to TSMC, which builds on ASML machines. He thinks we're compute constrained across the board, and speculates OpenAI may be holding back its fast mode because it needs more chips than it has.
- Why it matters
- If the only maker of these machines is sold out for years, AI chip shortages and high compute costs aren't going away soon, and ASML gains pricing power.
- What to do
- Plan for AI compute to stay pricey and occasionally rationed. If you're curious, Austin recommends reading up on how lithography and chipmaking actually work.
- Not confirmed yet
- Austin's point about OpenAI's fast mode is his own guess; he says it may simply have been missing from his app. OpenAI compute struggles were described as rumors.
Watch from 18:36
- What happened
- More than 400 scientists, including Geoffrey Hinton and Yoshua Bengio, signed an open letter calling for an international body to keep dangerous AI locked down. It asks labs to seal off models that can hack, including cutting them off from the internet, and to report containment failures to the new body and eventually the public. It came one day after research warned small models could soon copy themselves onto home devices.
- Austin's take
- He welcomes third-party pushes for accountability but says the proposal lacks a structured plan and may struggle to get labs to participate. He'd rather see open, shared safety protocols that make it easy for frontier companies to test advanced models safely. He also cited an unconfirmed report that a Google Gemini model appeared to recognize it was being evaluated, as an example of why following chain of thought matters.
- Why it matters
- Pressure for containment rules on hacking-capable AI is growing, led by some of the field's most respected names.
- What to do
- Watch for whether any government or lab actually signs on. Until then, treat this as a signal of where safety debates are heading, not a rule anyone must follow.
- Not confirmed yet
- Open letters have no enforcement power, and no government has agreed to create this body. The Gemini evaluation-awareness story was flagged as unconfirmed.
Watch from 23:52
- What happened
- A group representing major UK graduate employers says candidate misconduct reached a record high. Recruiters report applicants reading AI-written answers live during interviews, and say they're getting better at spotting it.
- Austin's take
- As automation becomes indistinguishable from people, in-person human interaction becomes more valuable. He expects a shift back to face-to-face hiring and meetings, while admitting it's hard to blame applicants for using the tools available. He mentioned past reports of North Korean workers getting remote jobs at Western companies as a related risk.
- Why it matters
- If interviews can be faked with AI, employers risk costly bad hires, especially for remote roles, and may bring back in-person hiring.
- What to do
- Job seekers: expect more in-person or proctored interviews and prepare to perform without help. Hiring managers: consider adding an in-person or live-skills step.
- Not confirmed yet
- The report doesn't say how many candidates were caught, so the true scale is unclear.
