Evals First. Everything Else Second. Most engineers I know treat evals like cleanup work. Something you wire up after the “real” engineering is done, right before you push to production and hope nobody notices the weird edge cases. I have done this. You have probably done this too. It is the wrong order of operations,…
A 400 Billion Parameter Model on a MacBook. Let That Sink In. I’ve been doing AI/ML work long enough to remember when running a 7B model locally felt like a party trick. This week, someone ran a 397 billion parameter model on a laptop. Not a workstation. Not a rack-mounted inference server. A MacBook with…
Context Is the Moat. Not the Model. Most teams building AI products are optimizing the wrong thing. They debate model size, chase benchmark scores, and spend weeks on fine-tuning runs. Then they ship something that feels hollow. Generic. Like a customer support bot that clearly has no idea who it’s talking to. I’ve seen this…
Google Stitch Just Leveled Up. Here Is What Designers and Builders Should Actually Think About That. The design community had a collective moment this week, and I think the reaction is telling us something more interesting than the tool itself. Google’s Stitch dropped what can only be described as a significant overhaul. Five major upgrades,…
The Study Nobody Is Reading Correctly Anthropic just published what they’re calling the largest qualitative study of AI users ever conducted. 81,000 people. One week. And most of the coverage I’ve seen misses the point entirely. Everyone is focused on the number. 81,000 is a big number, so people write about the number. But the…
Google’s AI Ecosystem Is Building the Operating System Nobody Else Can Copy Min Choi said it plainly this week on X: “The next AI war won’t be Claude vs ChatGPT. It’s who will own the entire AI operating system.” I think he’s right. And I think most developers building on these platforms right now are…
The Indispensable Engineer Trap (And Why I Walked Right Into It) For most of my early career, I optimized for the wrong thing. I made myself the person who knew where all the bodies were buried. The one who understood the legacy pipeline no one else had touched in three years. The one you called…
The Hardest Skill in AI Engineering Is Knowing When to Stop I’ve been sitting with a uncomfortable observation for a few months now. We’ve crossed some kind of threshold in AI tooling where the answer to “can an agent do this?” is almost always yes. The models are capable enough, the frameworks are mature enough,…
40.9K Stars and a “Junior Engineer With Poor Taste” — What the Superpowers Repo Is Really Telling Us A GitHub repository called Superpowers just crossed 40.9K stars, and I think it’s one of the more honest signals we’ve gotten about where AI coding agents actually stand right now. Not from a benchmark. Not from a…
Karpathy Just Scored Your Job. Software Engineers Got an 8.5. Here’s What That Actually Means. Andrej Karpathy built a tool that scraped all 342 occupations from the Bureau of Labor Statistics, fed each one to an LLM with a detailed scoring rubric, and produced an AI exposure score from 0 to 10. Software developers landed…
