Steve Yegge
@Steve_Yegge
I've been in the industry for O(40) years and have written O(1M) LOC. I don't think I'll ever write O(another) line again, but I'll be launching more than ever.
I came to the same conclusion as @michaelshuffett -- you hit a cognitive ceiling around 3-5 parallel workflows. Sometimes 2-3 will bog you down, and sometimes you can hit 4-5 streams at once, but either way you're always the bottleneck. You. Not them. And yet 95-99% of the time,…
Here's what I discovered after months of AI-assisted coding: @cursor_ai / Claude Code are insanely good... BUT managing multiple agents quickly becomes chaos. You hit a ceiling around 3-5 parallel workflows. Your brain just can't handle more.
Coding agents, it turns out, are both terrible at working with long functions, and great at generating long functions if you don't tell them not to. I've been in the habit of asking coding agents to review their code after generating it, in several passes. I have them check for…
I guess I can post this now that the dust has settled. So one of my favorite things to do is give my coding agents more and more permissions and freedom, just to see how far I can push their productivity without going too far off the rails. It's a delicate balance. I haven't…
Serious question: does anyone else besides me run their coding agents with permissions checks disabled? E.g. "claude --dangerously-skip-permissions" or putting "amp.commands.allowlist":["*"] in your ~/.config/amp/settings.json. I've been doing it this way for at least six or…
I've been running five @AmpCode coding agent instances for a couple months. One little piggy works on Emacs. One little piggy builds Node. One little piggy ports old tests. One little piggy reviews code. And one little piggy does wee wee wee random crap the other piggies can't be…
Everyone else is trying to figure out how to make their coding agent the cheapest. We only care about being the best.
Lol Amp just flat out beats the shit out of any other agent at the moment, to the point of being crushingly so. Goddamn.
I just put up a new blog post about the current rapid rise of Agentic vibe coding using autonomous agents. It's called "The Brute Squad": sourcegraph.com/blog/the-brute… This is my follow-up to Revenge of the Junior Developer. More jokes, more news, maybe some insights. I hope you find…
Excited that we're almost done writing our book. It's going out for peer review this week. It has been a beast to write, but turned out pretty well, I think.
We’re excited to unveil the cover (and final title) for the groundbreaking new book from industry titans @RealGeneKim and @Steve_Yegge, releasing this fall! Vibe Coding: Building Production-Grade Software With GenAI, Chat, Agents, and Beyond isn’t just another tech book—it’s the…
I almost want to hoard this for myself, but: ampcode.com/how-to-build-a… is amazing. @thorstenball deconstructs how the entire industry is pivoting to the same 350 lines of code. "There is no moat" -- pretty strong words. But I'm so excited. I want a native coding agent in…
Hi all, I just dropped a new blog post: sourcegraph.com/blog/revenge-o… This one's a beehive-kicker for sure. Hope you like it and find it enlightening, even if you don't agree with all of it.
Another late-night Claude Code post. First, if you've just arrived here at the party, Claude Code is NOT the same thing as Claude 3.7 Sonnet, Claude.ai, nor any other Claudey thing. It is its own new experimental thing, also from Anthropic, makers of Claude and…
