Finding signal on Twitter is more difficult than it used to be. We curate the best tweets on topics like AI, startups, and product development every weekday so you can focus on what matters.
PLUS: OpenClaw's 3,000 commit day
Supabase raised $500M at a $10B valuation and shipped passkeys, ChatGPT integration, and a Postgres operating system called Multigres.
OpenClaw shipped 2,700 commits in a single night during a refactor. The saving grace was overfitted unit tests AI loves to generate.
Linus Torvalds calls AI a revolution that changes nothing fundamental. Compilers were a productivity shift too, and nobody quit programming.
Software engineers at Meta are being assigned manual data labeling. Same thing happened at Scale AI under Alex Wang.
Run Hermes on your main computer, not a side machine. Desktop app plus background mode unlocks actual multi-agent workflows.
Franz founder fired his team of five and went back to solo after realizing committees slow shipping and balloon costs.
Trump says he's meeting with all the AI companies next week to discuss making the American public partners in AI success.
Elon is bringing in $2 billion a month for compute. Google pays $920M monthly, Anthropic pays $1.25B monthly for SpaceX capacity.
Building is cheap now. Early VC money is shifting from product to distribution as the real cost.
Sriram Krishnan is leaving the White House at the end of this month after shaping America's AI strategy.
Moore's law, reframed: every 18 months the value of computation halves.
Mythos will one-shot full stack mobile apps but cost $120 to $400 per app at API pricing, five times more expensive than Opus.
Former Meta L8 engineer now ships 40 PRs a day solo by running 20 to 30 agents in parallel with a system called No Mistakes.
AI writing is painful when every menu is filled with Claudisms. Reports are not what leaves the room.
You could build a top tier VC firm just on deep model benchmarking. Find capability overhang, track where models suck.
Gemini Pro hasn't released since February and the gap is widening. Flash doesn't close the distance to Claude or GPT.
Codex Mac app rocks but it's a sub-par designer compared to Opus 4.8 according to Andrew Wilkinson.
Paul Graham found a startup whose best secret weapon is genuine benevolence. Delightful when that's the most natural moat.
shadcn spent a week on Sidebar, several days on Input-group. Good abstractions take hundreds of examples and constant refining.
If pricing power fades at scale, capitalism as we know it is roughly over. Inputs going to zero breaks resource allocation.
People on Reddit argue over OpenClaw models like wine. Minimax is a daily driver, GPT-5 gets creative when it shouldn't.
AI got so good that one founder has way more customers but does less customer support and ships with fewer bugs.
Going on Starter Story means 100 people learn about you every hour for a year from a single video.