Good morning.

I want you to answer one question before you read any further: how much do you trust Sam Altman?

We’ve had multiple reasons not to in this newsletter. See here , here and here. We called him “Scam Altman” long before Elon Musk started doing it. And even before the (and this is not a joke) sexual assault and even murder allegations against him.

So - do YOU trust this guy?

Take your time. Because according to one of the longest, angriest pieces of tech journalism I've read this year, your answer is holding up the entire US stock market. Maybe the global economy.

Here's the argument, stripped to the bone (but feel free to read Ed Zitron’s 8000 word article).

The AI bubble is not an AI bubble. It's an OpenAI bubble. Remove OpenAI from the last five years and nothing inflates. No trillion dollars in data centers. No NVIDIA at the top of the world. No Anthropic either - Google and Amazon only funded Anthropic to have a horse against Microsoft's horse.

One popular website. That's what this entire edifice is standing on. ChatGPT got big fast, and the largest companies on earth decided that meant the future had arrived. They've spent over a trillion dollars on that belief.

And the numbers underneath it are ugly.

OpenAI plans to burn 852 billion dollars by 2030. It owes 748 billion in compute commitments to Microsoft, Amazon and Oracle. It has 900 million weekly users and roughly 5% of them pay. The rest cost money every time they type a prompt. The ad revenue that was supposed to fix this? Analysts think the entire AI chatbot ad market will be worth one billion this year. OpenAI alone projected 2.4 billion.

There's a phrase in the piece that stuck with me: hyperscaler psy-op. The idea that OpenAI and Anthropic were never really startups. They were willed into existence by Microsoft, Google and Amazon, who built their infrastructure, ate their losses, and pointed at their growth to justify spending hundreds of billions more. A closed loop. Money going in circles while everyone insists it's demand.

Now. Truth is - I don't fully buy the doom scenario that Zitron puts forward. I use these tools every single day. So do you, probably. Something real is happening underneath the financial madness. The author waves that away as irrelevant, and I think that's where the piece overreaches.

Ed Zitron hates AI so much he is getting quite angry in his editorials and sometimes I think it clouds his judgment.

But I cannot wave away the entire article based on that. He does have a point.

If OpenAI stumbles - misses a payment, breaches a contract, raises a down round - there is no plan B. The author calls it the Lehman Brothers of AI. Dramatic. Maybe not wrong.

And Anthropic doesn't escape this either. Same losses, same compute addiction, same hyperscaler life support. Just better PR and a slightly more trustworthy CEO.

I keep coming back to that opening question : can you trust Sam Altman. Not because Altman is a villain. Because we've built a system where one man's ability to keep raising money is the difference between a booming market and a crater. Nobody voted for that.

Maybe he pulls it off. Maybe the IPO lands in 2027 and this newsletter ages terribly. I'd honestly be relieved.

One company. One man. One story everyone agreed not to question.

So what about you? Do you trust him?

Welcome to the Blacklynx Brief

Apple sues OpenAI over stolen hardware secrets
Apple filed suit against OpenAI, hardware chief Tang Tan, and its io devices unit, alleging some of the 400+ employees OpenAI poached brought confidential hardware designs with them. Apple wants the unreleased Jony Ive-designed device sent back to the drawing board before its expected 2027 launch. (Court filing)

DeepMind's Hassabis pitches a US AI watchdog
Demis Hassabis published a plan for a FINRA-style US body that would screen frontier models for deception, bioweapons, and hacking capabilities 30 days before release. He wants it operational this year, warning open-source models could reach dangerous territory within 18 months. (Demis Hassabis on X)

OpenAI ships GPT-5.6 and takes aim at the workplace
OpenAI released the GPT-5.6 family, with flagship Sol landing just below Anthropic's Fable at a fraction of the price. The launch pairs with ChatGPT Work, a Codex-powered agent platform merged into a revamped desktop app with built-in browser and computer control. (OpenAI)

16 Nobel laureates put AI's job shock on the clock
Over 200 researchers and economists signed "We Must Act Now," a Stanford-organized statement warning AI could displace jobs at Industrial Revolution scale within a decade — but in years, not generations. Signatories include Google's Jeff Dean and Anthropic co-founder Jack Clark. (We Must Act Now)

New York freezes the AI data center boom
Governor Kathy Hochul ordered a moratorium of up to 12 months on new permits for data centers above 50 megawatts, the first statewide pause on the AI buildout. Regulators will use it to write water, air, and grid-cost standards, and Hochul wants the industry's tax breaks repealed. (Office of Gov. Hochul)

That’s it for us this week. Please like and subscribe 🙂