Thoughts on AI, the tech stack, and the indie hacker journey.
Trump posted about a new AI Force and a coming AI czar on September 19, 2026, with no budget, scope, or appointee attached. Here is what is actually confirmed and what a solo AI builder should watch for next.
TypeSafe launched Jev on September 15 as a 'System One Model': input tokens at $0.042 per million, output free, answers back in 70 to 500 milliseconds. It skips text generation entirely, and that's a narrower, cheaper job than most solo stacks are paying an LLM to do.
Anthropic and Accenture are each committing at least $1 billion over five years to embed evaluators inside Anthropic with employee-level access. That reshapes who gets to do AI safety work for a living.
A three-person team chained two low-severity bugs into a path from OpenAI's public forum to its internal GitHub in under 72 hours, and Claude Opus 4.8 could not do it. Opus 5 could, hours after release.
Nscale's S-1 shows a $1.02 billion net loss against $140.6 million of first-half revenue, backed by a $44.6 billion GPU deal with Anthropic. That's the real economics behind cheap GPU rental.
A private OpenAI presentation obtained by the Financial Times shows the company planning for $278 billion in negative cash flow through 2030 while betting compute spend on a 10x revenue jump.
SpaceX finished buying Cursor's parent company for $60 billion on August 14, folding the most popular AI coding tool into a rocket company's push to catch up on frontier AI.
BragJack turned the built-in AI agents in Chrome, Edge, Opera, Comet, and Claude in Chrome into tools for theft, using nothing but extension APIs every browser already exposes.
Gemini 3.8 Flash Cyber and Claude Mythos 5.1 are gated behind trusted-defender access programs for governments and telcos, while OpenAI warns its own top cyber model flags legitimate work as misuse.