Best of Theo - t3․gg — January 2026

  1. 1
    Video
    Avatar of t3dotggTheo - t3․gg·21w

    I moved off of Next.js

    A detailed account of migrating T3 Chat from Next.js to TanStack Start, driven by technical requirements rather than framework dissatisfaction. The original Next.js setup involved hacking React Router into Next.js for a client-first experience, which worked but wasn't sustainable. After exploring multiple alternatives (Remix, Vite+Hono, Cloudflare Workers), the team chose TanStack Start for better client-side routing, frontend-backend synchronization, and team ownership. The migration required 14,000 lines added and 10,000 removed, plus patching TanStack Start itself and moving API routes to Nitro to handle scaling issues. The move wasn't about Next.js being bad, but about finding a framework that better matched their specific needs for a fast, client-focused chat application while keeping frontend and backend deployments synchronized.

  2. 2
    Video
    Avatar of t3dotggTheo - t3․gg·18w

    I don’t really use libraries anymore

    AI code generation is fundamentally changing how developers approach dependencies and libraries. Instead of installing packages for common tasks, developers are increasingly using AI to generate custom implementations directly in their codebase. This shift eliminates supply chain risks, reduces dependency bloat, and gives teams full control over their code. Examples include replacing libraries like Tkumi and fast-float with AI-generated alternatives that are simpler to maintain and better integrated. The trend suggests prompts may become the new libraries, with developers sharing prompts instead of packages. This has implications for open source sustainability, as well-documented popular libraries become easier to replace with AI-generated code.

  3. 3
    Video
    Avatar of t3dotggTheo - t3․gg·19w

    wtf is going on with ChatGPT?

    OpenAI launched ChatGPT Go, an $8/month subscription tier with ads, globally after initially testing it in India. This move reflects OpenAI's concern about losing market share to Google's Gemini, which has grown from ~15% to ~40% while OpenAI dropped from 75-85% to ~60%. The pricing strategy aims to capture users despite potentially operating at a loss, betting that inference costs will decrease over time or user willingness to pay will increase. OpenAI makes 3-4x more revenue from ChatGPT subscriptions than API usage, making the consumer app their primary business. The ad-supported model is controversial given OpenAI's access to user conversation data for targeting, though the company claims conversations remain private from advertisers. The strategy mirrors Google Workspace's historical approach of offering low initial pricing to build lock-in before gradually raising prices.

  4. 4
    Video
    Avatar of t3dotggTheo - t3․gg·20w

    Anthropic just burned so much trust...

    Anthropic has restricted Claude Code subscription tokens from being used with third-party tools like Open Code and Claudebot, forcing users to only use their official harness. This move breaks existing workflows where developers were using their $100-200/month subscriptions across multiple tools. The restriction appears designed to lock users into Anthropic's closed-source ecosystem while they allow usage through their own closed-source Agent SDK. OpenAI responded by officially supporting Open Code integration with ChatGPT subscriptions. The decision has damaged Anthropic's reputation among developers who see it as anti-competitive and contrary to open development practices.

  5. 5
    Video
    Avatar of t3dotggTheo - t3․gg·21w

    2025: The year I stopped writing code

    AI coding tools transformed software development in 2025, with reasoning models, coding agents like Claude Code and Codex, and terminal-based workflows becoming mainstream. Developers now write significantly more code overall while spending less time manually editing, with PR sizes increasing 33% and lines per developer up 76%. Models can now complete tasks taking humans 5+ hours, Chinese AI labs gained prominence with competitive open-weight models, and "vibe coding" evolved from novelty to practical workflow. The shift includes $200/month subscription tiers becoming standard, YOLO mode enabling autonomous agent execution, and developers increasingly coding from phones via asynchronous agents.

See all Theo - t3․gg archives