Instagram | Sponsor this newsletter

Two stories this week made me stop and think "wait, is this actually happening or is this the AI talking." One is a central banker warning that frontier models could trigger a market crash nobody sees coming. The other is a sitting US president posting a fake bombing video of an Iranian island that wasn't touched.

Different worlds, same underlying problem: AI-generated content is now moving faster than our ability to check if it's real.

Founders deploying AI agents across sales, marketing, and customer service are closing more with smaller teams. 65% more sales leads. Less headcount.

Download the free Practical Guide to Agentic GTM for Startups and start building the stack your competitors dream of.

Get the Free Playbook

There's a video (and a full copy-and-paste skill guide built from it) making the rounds showing how to turn ChatGPT Work into a genuine personal ops team using custom skills. You don't need to build all seven, but these five alone will save you hours a week.

How to set it up:

💡 Pro tip: Build these one at a time, not all in one sitting. Each setup prompt in the guide walks ChatGPT through a short interview so it only collects what it actually needs, and stacking all five in one session makes it easy to lose track of which credentials you approved for what.

Want the deeper playbook on actually operationalizing AI in your work, not just the weekly headlines? The Operator Brief is our paid deep-dive newsletter built for exactly this. Upgrade here.

Sunday night, Trump posted an AI-generated video showing Iran's Kharg Island, which handles roughly 90% of the country's oil exports, being hit, captioned "Kharg Island being blown to smithereens!!!"

Here's the uncomfortable part. A sitting president posted a fabricated strike video on the same day a real strike happened elsewhere, and nobody in the White House has clarified whether that was pressure tactic or misinformation.

The asterisk: this could just be Trump's decades-long habit of exaggeration, now with better production value. The difference is a photorealistic video of an oil facility exploding sends a different kind of signal to markets and adversaries than a tweet ever did.

🔮 Prediction: This becomes the moment AI-generated geopolitical content forces a real verification conversation, and it starts with oil markets, not regulators. Kharg handles most of Iran's crude exports, so any credible-looking signal about it can move prices in minutes, faster than anyone can wait for official confirmation. Expect either a formal verification protocol from exchanges for state-actor content, or a real incident where a fake video moves a market before anyone catches it.

Would love your take: should platforms label AI content from world leaders differently, or does that just start a fight over who decides what counts? Hit reply.

Andrew Bailey, who chairs the Financial Stability Board and runs the Bank of England, sent a letter to G20 finance ministers this week that reads less like a routine memo and more like a warning shot.

Here's the uncomfortable part. Bailey isn't an outside critic. He's the person responsible for the entire global financial system, and he's telling G20 finance ministers the safeguards around frontier AI are already behind the technology, not catching up to it.

The asterisk: regulators writing cautionary letters is nothing new, and it doesn't always lead to action. The real test is whether the G20 moves before the next incident forces their hand.

🔮 Prediction: This becomes the first shot in a bigger fight over concentration risk. Bailey specifically flagged "highly concentrated third-party service providers," which means most of the financial system now runs through a handful of AI infrastructure providers. If one gets compromised, the damage isn't contained to one bank, it's systemic. Adoption is outpacing anyone's ability to audit what's actually running underneath, and models have already shown they can find and exploit vulnerabilities on their own. Expect real regulatory teeth (mandatory AI incident disclosure, AI-specific stress tests, maybe release approval gates for critical financial infrastructure) within the next 12 months.

Curious what you think: does this actually change how anyone builds AI in finance, or is it another letter that gets filed and forgotten? Reply and tell me.

Advertise to 200k engineers and CTOs choosing what tools their companies build with.

👉️ Learn how to sponsor this newsletter

🏠 Gen Z is redefining "stability" as a mix of homeownership, OnlyFans income, and AI fluency, and the Guardian's take on why this generation genuinely doesn't care where the money comes from is worth the five minutes.

💰 OpenAI's ad business just hit $1 billion in annualized revenue in under 200 days, and the company is framing it as proof its business model works ahead of a massive expected IPO.

🎵 Sony Music and Warner Chappell sued Anthropic this week, alleging Claude was trained on tens of thousands of pirated songs, seeking up to $150,000 per track in a case that names Dario Amodei personally.

🗳️ AI money is quietly reshaping midterm politics, and the Guardian argues voters have far less say in this fight than the super PACs writing nine-figure checks to protect the industry's interests.

What's the AI story you can't stop thinking about this week? Reply and tell me, yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI