Practical AI skills for people whose job isn't AI. Most posts end with a drill, because reading about a skill isn't having one.
The pieces that hold up regardless of what shipped this week.
New York signed a chatbot child-safety law, OpenAI rewrote its Model Spec, and Anthropic's usage policy changes take effect November 12. Three rule changes in one week — check whether your team's AI policy accounts for any of them.
Four separate 2026 agent failures passed review before anyone caught them, because every review checked the final answer instead of the actions that produced it. Here's the layer that actually catches problems, and a twenty-minute way to check it this week.
4 min read
Andrej Karpathy built a complete product in about two hours using an AI agent. Simon Willison shipped a playable game through 321 agent commits — and both point to the same shift in what separates a good result from a mediocre one.
4 min read
Three product changes this year quietly removed the editing pass between AI output and the finished file. Here's what has to move into the prompt now that nothing catches a gap afterward.
4 min read
Four frontier-class models launched within days of each other, each strongest at a different kind of work. Here's how to match model to task without re-running the comparison every month.
4 min read
AI now produces more output, in more formats, and faster than one person can read start to finish. Four signals from the past six weeks show exactly where that habit breaks — and the checking habit that replaces it without slowing you down.
4 min read
Anthropic interpretability researcher Emmanuel Ameisen told Lenny Rachitsky that AI models are grown, not built, and plan several steps ahead using reasoning that isn't human-like. Here's what that means for the next confident answer an AI gives you.
4 min read
OpenAI's CEO says AI security can't be patched into place. Anthropic's own IPO filing shows the business burning cash faster than it grows. Plan for both admissions instead of assuming they're already resolved.
4 min read
Anthropic's draft IPO filing shows $16.3 billion in first-half revenue against an $80.7 billion net loss and $518 billion in compute commitments. The numbers are a vendor-risk planning problem for any team paying for AI, not just a headline about a funding round.
4 min read
Most AI pilots don't fail loudly — they stop getting mentioned in standup while the subscription keeps renewing. Three questions that force the decision nobody schedules.
4 min read
Practical AI skills, sent when there's something genuinely useful to share. No filler, no hype.
We use one cookie to keep you signed in — that one isn't optional. Separately, we'd like to record which drills people finish and where they get stuck, so we can fix the parts that don't teach. No ad networks, no selling anything on. Privacy policy.