Julio
172 posts
Julio
@bolt_intern
Senior Full Stack Engineer at Bolt, part time intern and X shitposter. Previously technical founder and CTO in cryptoland
参加日 Kasım 2025
93 フォロー中30 フォロワー
so far, in the next version of Bun:
- global bun install virtual store reduces disk usage
- HTTP/3 Bun.serve() & fetch
- HTTP/2 fetch
- node:tls uses less ram
- Stability fixes for Worker, MessagePort & BroadcastChannel
guess the 2 things we’ll add to this list before we ship
English
@justinsunyt @capydotai Does this work only with vms or can i use it fully local? looks amazing
English
if anyone is looking for a better way to orchestrate parallel coding agents try @capydotai
captain orchestrates build tasks by default with native PR stacking primitives and works end to end so all you have to review are merge-ready PRs
English
kanban doesn’t make sense for coding agents
we tried it 6 months ago. every task just ended up in the “needs review” column
Muhammad Zahid@mzahidtech
🚀 This is wild. @cursor_ai just dropped a Linear-style Kanban board where you can literally drag in tasks and Cloud Agents pick them up, work on them, open PRs, and ship results. Built with the new Cursor SDK. Full agent orchestration in one dashboard. Mind officially blown.→ Check the example: github.com/cursor/cookboo… #Cursor #AIagents
English
@justinsunyt don't you guys have the kanban in the landing page though
English
I find most of the Anthropic hate to be wildly exaggerated but this is an area that will make or break them.
dax@thdxr
the whole "cannot use our models to develop a competing product" is getting really confusing now that model labs are launching regular products like Claude Design
English
@DrizzleORM Nice now we just need Effect v4 to actually ship their docs
English
Drizzle v1.0.0-rc.1 is out 🚀
▪︎ Effect v4 native support
▪︎ JIT row mappers to reduce ORM overhead to ~0
▪︎ Reworked casing API (breaking change)
▪︎ Drizzle for LLM agents (preview)
Drizzle is now as fast as using raw driver and mapping(or not mapping) results by hand 🙃
English
✅ Best harness
✅ Best UX
✅ All frontier models available
Cursor is the best place to write code with agents, full stop.
Lenny Rachitsky@lennysan
Narrative violation: Finding it's more fun to work within Cursor than in the native Codex or Claude Code apps. Not a massive difference, but just enough to keep me there. And obviously easier to play with new competing models as they come out. Nice work, Cursor.
English
@fireblaze_15 @Cloudflare Can me agent one shot delete everything though? I like to take some daily risks
English
@threepointone I have to build something where our internal team can build custom agents with code execution and mcp and I think I just need to wait until you finish this so I can fork it
English
been working on subagents etc for the last few days, and it's come along well. some highlights:
(tldr - we have the world's best subagents. they're just baby durable objects.)
- subagents as separate threads so you can host all of a user's/entity's chats in a single DO (examples/multi-ai-chat)
- subagents now have a first class story as "tools", with resumability, recovery, persistence etc all built in. works with ai-chat and think. (examples/agents-as-tools)
- a new example showing how to use github auth for your agents (because I want to actually deploy something live for my friends and I to use. examples/auth-agent)
- can now configure beforeStep, reasoning, telemetry for think.
- fixed a few nasty timing/consistency bugs
just did a release. now starting a big docs sync for everything we shipped in the last month, aiming for it all to go out tomorrow. we should update the starter with the new fun stuff as well sometime...
npm i agents
English
@threepointone Oh that would unlock so much like on prem enterprise deployments
English
@bolt_intern I'm going to fix our self hosting story and make that a non issue. hopefully this quarter.
English
vendor lock-in is a dead concept in a world where your coding agent can rewrite the entire stack in under an hour. cloudflare's (and every other provider's) only bet is to provide good services and a good experience for a good price.
𝗺𝗮𝗹𝗶𝗽𝗲𝘁𝗲𝗸 🇹🇷🇵🇸 goyim in abundance@malipetek
@threepointone Cloudflare just needs one thing to explode imho: vendor lock-in insurance.
English
@threepointone @michellechen I really hope you get that head of forks promotion Sunil
English
you can only eat pi with forks
Mario Zechner@badlogicgames
People of pi.dev Cloudflare Workers AI release. We want to see forking company and triangle company battle it out to the death. So since we had Vercel Gateway in pi for quite a while, we now have merged a PR that adds CF Workers AI as well. CF AI Gateway yet to come.
English
@zeeg @snellingio Opus 4.7 loads so much context so fast for some reason. Is often at over 250k context for me for normal sessions
English
i don’t understand what you gain moving 4.6… it’s not as if 4.6 is cheaper?
opus money is opus money
David Cramer@zeeg
just asked the team to turn off opus 4.7 no true gains in performance, burning compute (and money) honestly, imo, a bad release from Anthropic
English
After the Cursor + xAI deal, there are almost no major IDE vendors left that aren’t backed by Big Tech cloud.
Welcome to the new money funnel for devs:
Antigravity → Google
Kiro → AWS
Copilot + VS Code → Microsoft
Cursor → xAI
Almost.
Kirill Skrygan@kskrygan
Would you be interested if JetBrains releases a totally local AI agent, working 100% on your laptop, using our code insight engine and deeply integrated into the IDE? Yes, it will be probably 1 month behind the very recent frontier models, but no token blood bath anymore WDYT?
English
@samgoodwin89 We need an effect version of project think so I can drown in the koolaid
English
About to present on Alchemy v2 at the Effect office hours. Come check it out.
Effect | TypeScript at Scale@EffectTS_
-6 hours! 🔥
English
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
i hate flakey tests
English
@dillon_mulroy Lol I was literally just watching your conf from effect days 2025 and was like: isn’t this a orange cloud yapper? Haha
English
effect on cloudflare is so back.
already can tell alchemy will be the way i build
Effect | TypeScript at Scale@EffectTS_
🔴 We're live on YouTube! youtube.com/watch?v=fPSxB7…
English
paste the SDK docs to @cursor_ai and build your dream agent orchestration with:
- any model
- built-in cloud agents support
- anywhere you want
exciting! more to come :)
Cursor@cursor_ai
We’re introducing the Cursor SDK so you can build agents with the same runtime, harness, and models that power Cursor. Run agents from CI/CD pipelines, create automations for end-to-end workflows, or embed agents directly inside your products.
English