Best of Theo - t3․ggApril 2026

  1. 1
    Video
    Avatar of t3dotggTheo - t3․gg·5w

    GitHub has a fake star problem…

    A deep dive into the fake GitHub star economy, covering a peer-reviewed CMU/NC State/Socket study that identified ~6 million suspected fake stars across 18,600 repos. The investigation reveals a mature shadow market where stars sell for 3 cents to 90 cents each, with ROI of up to 117,000x when used to manufacture VC funding traction. Key findings include detection heuristics like fork-to-star and watcher-to-star ratios, analysis of specific AI and blockchain repos showing manipulation signals, how VC firms like Redpoint explicitly use star counts as sourcing benchmarks, and GitHub's reactive-only enforcement leaving the fake account infrastructure largely intact. The piece also covers legal exposure under FTC rules and SEC wire fraud precedents, and recommends alternative metrics like contributor activity and package downloads.

  2. 2
    Video
    Avatar of t3dotggTheo - t3․gg·6w

    Did Anthropic just kill Figma?

    Anthropic launched Claude Design, a new AI-powered UI prototyping tool that competes directly with Figma. The author does a live first-look, using it to redesign the T3 Code marketing site. Key features include importing a codebase for context, leaving comments on designs for batch AI edits, a 'knobs' mode for live CSS tweaking, and exporting designs to Claude Code for implementation. The tool shows genuine promise for rapid UI mockups and bridging the designer-developer gap, but suffers from usage limits (separate quota from regular Claude), bugs like disappearing files, and rough edges. Despite issues, the author concludes it's the best software Anthropic has shipped and a real threat to Figma's market position.

  3. 3
    Video
    Avatar of t3dotggTheo - t3․gg·7w

    I think every company should open source their code.

    A strong argument for why companies should open source their software, framed around the emerging 'building block economy.' The core thesis is that AI agents prefer open, modular, well-documented components over closed commercial software, and that the future of competitive advantage lies in letting customers fork and customize your product rather than building every feature yourself. Uses T3 Code's 1,500 forks and Mitchell Hashimoto's Ghosty/libghosty growth data as evidence. Also introduces the concept of a 'patch.md' file — a plain-English description of user customizations that enables AI-assisted merge conflict resolution when upstream updates break forks — as a path toward self-forking, self-healing software.

  4. 4
    Video
    Avatar of t3dotggTheo - t3․gg·8w

    Bash is bad for agents

    Bash is a useful but insufficient execution layer for AI coding agents. The core argument is that flooding LLM context windows with code (e.g., via tools like Repomix) degrades output quality and increases cost, while letting models use bash to fetch only relevant context is far better. However, bash lacks standards for destructive vs. safe operations, permission management, portability, and multi-user isolation. TypeScript/JavaScript is proposed as a superior alternative execution environment because it supports typed inputs/outputs, runs in isolated V8/Node environments without full VMs, enables deterministic filtering, and reduces token usage. Projects like Cloudflare's Code Mode, Vercel's JsBash, and open-source executor experiments are highlighted as early steps toward a proper agent execution layer.