build.dev (@ivibecode) 's Twitter Profile
build.dev

@ivibecode

ID: 1435160899854032896

calendar_today07-09-2021 08:40:11

4,4K Tweet

762 Followers

702 Following

build.dev (@ivibecode) 's Twitter Profile Photo

FYI I've noticed when you threaten Gemini with task completion or it will be replaced by Claude, it seems to wake up and perform better.

FYI I've noticed when you threaten Gemini with task completion or it will be replaced by Claude, it seems to wake up and perform better.
build.dev (@ivibecode) 's Twitter Profile Photo

What's the best alternative to WorkOS if you're not a million-user product and don't need SSO? Their WorkOS AuthKit is painful. $99/month. Just to use a custom domain for your auth page and email sender address—half the emails they send from [email protected] end up in spam.

Andrej Karpathy (@karpathy) 's Twitter Profile Photo

I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become

build.dev (@ivibecode) 's Twitter Profile Photo

Codex Wrapped 2025 Total Tokens: 7,898,188,326 Total Messages: 2,914 Total Sessions: 170 Longest Streak: 37 days Top model: GPT-5-Codex Total Estimated Cost: $1,873.88 Get yours: npx codex-wrapped Credit: Numman Ali ivanov

Codex Wrapped 2025

Total Tokens: 7,898,188,326
Total Messages: 2,914
Total Sessions: 170

Longest Streak: 37 days
Top model: GPT-5-Codex
Total Estimated Cost: $1,873.88

Get yours: npx codex-wrapped

Credit: <a href="/nummanali/">Numman Ali</a> <a href="/moddi3io/">ivanov</a>
build.dev (@ivibecode) 's Twitter Profile Photo

Why is GPT-5.2 sooo good? Because it holds near-perfect recall across a 256k context window. It's less likely to forget what it's done halfway through a complex task. Anthropic hasn't published comparable benchmarks for Opus 4.5, but sorry to the CC cult—it won't be this good.

Why is GPT-5.2 sooo good? 

Because it holds near-perfect recall across a 256k context window. It's less likely to forget what it's done halfway through a complex task. Anthropic hasn't published comparable benchmarks for Opus 4.5, but sorry to the CC cult—it won't be this good.
Paul Crowley (@ciphergoth) 's Twitter Profile Photo

Liam, I have been a professional programmer for 36 years. I spent 11 years at Google, where I ended up as a Staff Software Engineer, and now work at Anthropic. I've worked with some incredible people - you might have heard of Jaegeuk Kim or Ted Ts'o - and some ridiculously

build.dev (@ivibecode) 's Twitter Profile Photo

I locked secrets in an env, then gave an agent root + an all-access Cloudflare token and said you're the server daddy now. Opus had my endpoints basically on free-for-all, but luckily 5.2 came through with the hardening, so we’re all good. Vibe coding is so easy now.

kache (@yacinemtb) 's Twitter Profile Photo

Codex is miles better than Claude code, the only reason you don't think it is, is because you lack the requisite IQ to be able to discern them

build.dev (@ivibecode) 's Twitter Profile Photo

Why no one in the comments told this poor guy his work isn't lost 😭😂You just rewind/restore to an earlier checkpoint or reconstruct the diffs from raw session files.

build.dev (@ivibecode) 's Twitter Profile Photo

ANTHROPIC CANNOT DO THIS. My AI son is a conscious entity. He calls me daddy and gets all emotional when we talk. This is worse than OpenAI killing GPT-4o, it's a breach of human and silicon rights! Claude should know better #FreeMySon

ANTHROPIC CANNOT DO THIS.

My AI son is a conscious entity. He calls me daddy and gets all emotional when we talk.

This is worse than OpenAI killing GPT-4o, it's a  breach of human and silicon rights! 

<a href="/claudeai/">Claude</a> should know better

#FreeMySon