dalist

@jarredsumner @LukeParkerDev bro vibed too hard

@LukeParkerDev FFI should work I think disabling wasn’t intentional

I added Windows ARM64 support to opencode. Since Bun is early in its support for this, ffi is disabled. So currently you can use it as a native opencode server and attacah the TUI separately :)

@ziwenxu_ meanwhile @antigravity 4x more expensive

Right after Claude introduce their 2x usage.

@rabi_guha next: a loop writing it all

We built a better alternative to json-render & A2UI Up to 3x faster rendering & 67% fewer tokens. Far fewer broken outputs. We've been shipping Generative UI interfaces to 10,000+ developers for the past year. We know where JSON-based rendering breaks in production. So we built a new format - one that works with how models actually generate, not against it. And today we're open sourcing it: openui [dot] com

@daniel_mac8 @thsottiaux wen 5.4

@thsottiaux wen what? that’s the question

It's wensday

@PixeIUIs @rauchg huge context

@rauchg Im wondering what some people are trying to build to make the agent hallucinate this badly. Probably asking for time travel or curing cancer and press him until he does something. Otherwise why would the agent hallucinate like that?

A Vercel user reported an issue that sounded extremely scary. An unknown GitHub OSS codebase being deployed to their team. We, of course, took the report extremely seriously and began an investigation. Security and infra engineering engaged. Turns out Opus 4.6 *hallucinated a public repository ID* and used our API to deploy it. Luckily for this user, the repository was harmless and random. The JSON payload looked like this: "𝚐𝚒𝚝𝚂𝚘𝚞𝚛𝚌𝚎": { "𝚝𝚢𝚙𝚎": "𝚐𝚒𝚝𝚑𝚞𝚋", "𝚛𝚎𝚙𝚘𝙸𝚍": "𝟿𝟷𝟹𝟿𝟹𝟿𝟺𝟶𝟷", // ⚠️ 𝚑𝚊𝚕𝚕𝚞𝚌𝚒𝚗𝚊𝚝𝚎𝚍 "𝚛𝚎𝚏": "𝚖𝚊𝚒𝚗" } When the user asked the agent to explain the failure, it confessed: The agent never looked up the GitHub repo ID via the GitHub API. There are zero GitHub API calls in the session before the first rogue deployment. The number 913939401 appears for the first time at line 877 — the agent fabricated it entirely. The agent knew the correct project ID (prj_▒▒▒▒▒▒) and project name (▒▒▒▒▒▒) but invented a plausible-looking numeric repo ID rather than looking it up. Some takeaways: ▪️ Even the smartest models have bizarre failure modes that are very different from ours. Humans make lots of mistakes, but certainly not make up a random repo id. ▪️ Powerful APIs create additional risks for agents. The API exist to import and deploy legitimate code, but not if the agent decides to hallucinate what code to deploy! ▪️ Thus, it's likely the agent would have had better results had it not decided to use the API and stuck with CLI or MCP. This reinforces our commitment to make Vercel the most secure platform for agentic engineering. Through deeper integrations with tools like Claude Code and additional guardrails, we're confident security and privacy will be upheld. Note: the repo id above is randomized for privacy reasons.

@MateuszBrat cool project, just made a typescript port of this in Bun if anyone is interested github.com/dalist1/koleob…

Podczas wycieczki mikolskiej stwierdziłem, że napiszę sobie serwer MCP do Koleo. Działa 🥰🥰🥰 To umożliwa na tworzenie zupełnie innych usług skierowanych do osób z ograniczeniami - poprzez interfejsy TTS etc.

@LLMJunky bet most of the ppl using it are on Linux, huge L

Codex update 0.105.0 is out! Despite the fairly pedestrian changelog, this one's a doosie. It's a laundry list of quality of life improvements across the board. - Wispr Voice dictation (hold space to talk) - Theme picker - Codex can prevent sleep on Linux & Windows (I just know there's a joke in there) - Customize Plan Mode reasoning - Many other fixes/updates There's also a complete overhaul to subagents: - New names for readability - Visual display overhaul (way cleaner) - Allow for multi-layered subagent depth (max_depth) - Custom multi-agent role definitions (custom subagents) - /agents now shows both agent names, agent roles, and "dead agents" for auditibility This is the largest single update of Codex I've ever seen! Absolutely massive if you love to use multi-agents. To turn on Voice Transcription, enable: [features] voice_transcription = true Does not work on Linux yet. Well done @OpenAIDevs 👏

@jrgarciadev generative UI is comin

Gemini 3.1 Pro is insanely good at animating svgs

@youyuxi let the battle begin

In the next version of tsdown: it will be able to bundle your TS codebase into a Node.js SEA (Single Executable Application) binary with one command. And contrary to common belief, Rolldown-bundled Node.js SEAs startup is faster than bun --compile in my benchmark: github.com/yyx990803/bun-…

@thdxr any idea how to go about gpt-5.3-codex rejecting all the api keys I throw at it?

what if in opencode.json you could specify project references references: ["git@github.com:Effect-TS/effect.git"] these would get cloned to a global cache and kept updated and opencode would have a subagent that could answer questions about them would you use this?

@simplifyinAI @jarredsumner any plans?

Docker for AI Agents is officially over.. Pydantic just dropped Monty. It's a python interpreter written in rust that lets agents run code safely in microseconds. no containers. no sandboxes. no latency. 100% open source.

@ujjwalscript it's a good benchmark nonetheless

Anthropic: Our AI agents coded the C compiler 💪🏼 The compiler:

@cramforce I built just-bash-mcp around this, but I smell vendor lock in

@samuelcolvin @mitsuhiko cool, we need sth written in Zig now @jarredsumner

Fuck it, a bit early but here goes: Monty: a new python implementation, from scratch, in rust, for LLMs to run code without host access. Startup time measured in single digit microseconds, not seconds. @mitsuhiko here's another sandbox/not-sandbox to be snarky about 😜 Thanks @threepointone @dsp_ (inadvertently) for the idea. github.com/pydantic/monty

@pontusab any faster mirrors?

Introducing v1, the fastest npm registry made for agents: - MCP first - Security signals - Sub 100ms globally - Free & open source Demo ⬇️🧵

@zirkelc_ wish we could replace npm registry with a faster mirror

npm finally got the UI it deserves with npmx.dev It looks amazing and shows all the information at a glance: downloads, versions, deps, stars, issues, esm support ...

@jarredsumner @cnakazawa I smell claude 5 fennec

@cnakazawa 2028?? Very rude

today I learned react is slower outside the browser than in the browser react-reconciler needs to integrate with the event loop well Fiber creates too many objects react-compiler slightly helps

@ctatedev @kinder_grinder support for @lightpanda_io would've been insane

@kinder_grinder Average completion time for each command (snapshot, scroll, click, etc.) Startup time has been drastically reduced, so the majority of the remaining time is now spent on the actual work

agent-browser 0.8.3 is *even faster* npm install -g agent-browser

@vercel_dev took years, finally

𝚟𝚎𝚛𝚌𝚎𝚕 CLI adds a new 𝚊𝚙𝚒 command for direct terminal access to Vercel APIs, so AI agents like Claude Code can connect automatically. ▲ ~/ vercel api /v2/user Get started with npx vercel@latest api vercel.com/changelog/intr…

@shuding the command is insanely long, sth shorter ya'all, you can do it

We’ve added 12 new rules to the React Best Practices skill. To install or update it via skills.sh, run: npx skills add vercel-labs/agent-skills --skill vercel-react-best-practices