This security panic looks very convenient for Anthropic’s IPO. That’s my hypothesis. In issue 09, I unpack PHP2Shell, Railway, and eMudhra, examine the damage we can actually verify, and explain why the commercial incentives caught my attention.

In this issue, I ask whether Anthropic’s IPO will be AI’s Facebook moment or its Lehman moment.

AI can build the product. It cannot make anyone care. @a16z with its 'New Media' understood what comes next faster than others. Media becomes startup infrastructure as important as data centers. My new newsletter issue is about why.

In Issue 06 of my newsletter, Tightly Coupled, I explore how Palantir created the FDE model, why everyone in AI is adopting it, and what happens to implementation, pricing, product discovery, and software economics when products begin deploying themselves.

If software, science, and robotics become nearly free to produce, what happens to labor, power, geopolitics, and to us? •_•

Another Issue of my newsletter: Cursor made “use the model’s code in one click” feel natural. The next step is software that understands what should change and builds it.

Everyone is suddenly building an “AI OS.” But are ChatGPT and Claude really becoming operating systems? In this issue, I explore what would actually need to change for AI to become the layer everything else runs on.

The second issue of my newsletter is out. Goes perfectly with your Sunday espresso ☕ The invisible market, the budget moving toward products agents prefer, and unknown unknowns.