Podbit · Lex Fridman Podcast
DHH runs Claude Code in tmux via a tool called Herder across 4-5 networked machines connected by Tailscale and GL.iNet Comet KVMs. He manages ~16 parallel agent threads, uses NeoVim as a project browser and LazyGit for changelogs, and orchestrates everything on Linux. Each machine runs 3 agents simultaneously.
Where this was said
At 1:46:06 · chapter starts 1:33:48
The tone shifts here into something more personal. DHH acknowledges that coming into the Linux community as a two-and-a-half-year newcomer draws legitimate pushback from people who've been doing it the hard way for decades. His equanimity about critics — 'every time there's a new flare-up, it's less' — reflects genuine experience with controversy across his career. The three-week X detox this summer came when his feed had somehow tilted toward politics and away from tech, even when he agreed with the political angles being presented. The algorithm holds up a mirror to your revealed preferences, he notes, and those aren't always pretty — Jung's shadow territory. His meta-observation about X actually improving because Bluesky and Mastodon became honeypots for the most aggressively tribal users is counterintuitive but backed by his personal experience.
When ATMs arrived, bank tellers feared for their jobs. Instead, branch costs fell, banks opened more branches, and teller counts increased. The same Jevons Paradox might apply to programming: as AI dramatically lowers the cost of software, demand for software could explode and create more developer opportunities — though DHH admits this isn't guaranteed.
DHH runs up to 16 simultaneous agent threads across 4-5 machines networked via Tailscale, representing his current maximum productive capacity.
Sam had no coding knowledge, so he used ChatGPT voice mode to generate his entire codebase and copy-pasted it into Notepad. A friend later introduced him to Cursor, and he never looked back.
Copy days of Discord chat history, paste it into ChatGPT, and ask it to list recurring pain points. The ones that come up most often are your best product bets.
Sam's top advice: when prompting Cursor, tell it to architect code for 100,000 users from day one. The AI changes its approach, building scalable frameworks instead of brittle one-user code.
With AI coding tools like Cursor, Bhanu replicates an existing free tool for a new keyword in under 5 minutes. What used to be a multi-day build is now a lunch-break task.
Ahrefs, SiteGPT, Cal.com, PostHog, Datafast, Sibyl AI, Bento, Feather, Featurepace, Mintlify, Cloud Code, ChartMogul — Bhanu runs his entire business solo with these 12 tools.
Frontier AI labs like OpenAI and Anthropic are priced as if they'll dominate every market while everyone else loses money. That's not how markets work, and cheap open-weight models are already proving it.
The best business model in history is offering 80% of the premium product at 50% of the price — Old Navy did it to Gap, Southwest did it to Delta, and now Chinese open-weight models are doing it to OpenAI. The value proposition is brutal.
China has every incentive to subsidize cheap AI to undercut US frontier labs. If OpenAI and Anthropic announce a growth slowdown, 40% of the S&P 500 gets torched — and that might be the point.
The price gap between open-weight and frontier AI models is staggering. Deepseek v4 Pro runs 30 cents per task. Claude Opus 4.5 runs $12.50. That's a 40x difference — and the UK's AI Security Institute found the performance gap is narrowing.
We use essential and analytics cookies to run Vuci. To understand how the site is used: Privacy Policy.
Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.