OK, change my mind:
$20 OpenAI subs for code review and planning
$10 OpenCode Go subs for orchestration and subagents
$7/week Fireworks Fire Pass for coding
~$60/mo is more than enough for 90% of projects.
You don't need Opus 4.6 with max reasoning for a vibe-coded Next.js SaaS.
GLM-5 is more than enough to work with. It can happily replace Opus 4.6, and the communication is better than GPT-5.4.
Kimi K2.5 with Fire Pass almost does the same job as Sonnet 4.6.
Yes, GLM-5 and Kimi K2.5 hallucinate, but so do Claude models. I've been working heavily with both Claude and OpenAI frontier models since August 2025, and I can easily replace them with cheaper models.
If you're working on vibe-coded or OSS projects, ZDR isn't really your problem, imo. If ZDR matters to you, go with OpenCode Go or Zen.
If you need both ZDR and speed, go with Fireworks. 100+ tok/sec for both GLM-5 and Kimi K2.5.
I can't believe people are building shitty apps and paying for multiple $200 subs on this platform. You don't need that.
You should define project rules and constraints in a testable way. Agents will hallucinate anyway. Telling them "don't do this" doesn't work. But if you enforce rules via lint or AST checks, they hit failures, see what's wrong, fix it, and move on.
My dear front-end developers (and anyone who’s interested in the future of interfaces):
I have crawled through depths of hell to bring you, for the foreseeable years, one of the more important foundational pieces of UI engineering (if not in implementation then certainly at least in concept):
Fast, accurate and comprehensive userland text measurement algorithm in pure TypeScript, usable for laying out entire web pages without CSS, bypassing DOM measurements and reflow
@thunkoid I'm building Mini Progress(iOS app)
Track reading goals, gym sessions, savings — right from your home screen widget. No app opening needed.
https://t.co/4lbesHOiP3
Build with Nansen CLI. Win a 𝐌𝐚𝐜 𝐌𝐢𝐧𝐢 𝐌𝟒 (16GB RAM, 256GB SSD)
4 weeks. 4 Mac Mini M4s per week. Weekly winners.
This week's prizes:
🥇 1st place: Mac Mini M4
🥈 2nd place: 50,000 Nansen API credits
🥉 3rd place: 10,000 Nansen API credits
Install and use our CLI, share it on X, and you're in.
Here's how to be eligible 👇
🚨 RIP Chrome for AI agents.
Someone built a headless browser from scratch that runs 11x faster and uses 9x less memory.
It's called Lightpanda.
Every AI agent doing web automation right now is running Chrome under the hood. That means you're spinning up a massive desktop application, stripping out the UI, and running hundreds of instances of it on a server. For something that never needs to render a single pixel.
It's like renting a semi-truck to deliver a letter.
Lightpanda is built differently. Not a fork of Chromium, Blink, or WebKit. Written from scratch in Zig with one goal: headless performance, nothing else.
It still runs JavaScript. Still handles Ajax, XHR, Fetch, SPAs, infinite scroll, all of it. Just without dragging along 500MB of browser bloat you'll never use.
And it drops straight into your existing stack:
→ Compatible with Playwright, Puppeteer, and chromedp via CDP
→ One-line Docker install
→ CDP server on port 9222, swap it in for Chrome in 30 seconds
The use cases are obvious: AI web agents, LLM training data scraping, browser automation at scale, testing pipelines. Anything where you're paying for Chrome compute and cringing at the bill.
It's still in beta and Web API coverage is growing. But at 11.8K stars it's clearly hitting a real nerve.
100% Opensource. AGPL-3.0.
Link in comments.
After chatting with many teams about their development and the current state of their projects, the next @HyperliquidX House of all Finance paper is finally here.
(https://t.co/jqH8HwcyG4)
🚨Claude Opus 4.6 wrote vulnerable code, leading to a smart contract exploit with $1.78M loss
cbETH asset's price was set to $1.12 instead of ~$2,200. The PRs of the project show commits were co-authored by Claude - Is this the first hack of vibe-coded Solidity code?