
Mark
36 posts

Mark
@markstachowski
Problem solver first, developer second. I simplify until the answer reveals itself.
Присоединился Ağustos 2025
447 Подписки0 Подписчики


i love having access to frontier models, they absolutely cooked on this one
OpenAI@OpenAI
GPT-5.5 delivers this step up in intelligence without compromising on speed. GPT-5.5 matches GPT-5.4 per-token latency in real-world serving, while performing better across nearly every evaluation we measured. It also uses significantly fewer tokens to complete the same Codex tasks, making it more efficient as well as more capable.
English


@petergyang They do this to us on purpose, AI drug dealers giving us our fix 😂
English

@romainhuet Couldn't be more obvious @romainhuet GPT-5.45. Didn't even use AI for this.
English

@thsottiaux Try to keep the Codex windows versions in feature parity. You guys are doing great things. Keep cooking.
English

@petergyang They don't nerf the models, they nerf all the harness logic around it. Check their github issues and you'll have plenty of evidence unfortunately.
English

If you think that’s good, just wait till you see what the team’s cooking.
Riley Brown@rileybrown
Codex App > Claude Desktop App
English

Please remove me from the OSS program @bcherny @AnthropicAI. I've tried contact support for weeks. I'd rather pay for the $200 plan like I have been.
For anyone who doesn't know, the OSS Program gives a $200 a month max plan for free for 6 months, but it's a gimped version of the max plan, so I haven't been able to use a large majority of new features because I have this OSS flag.
English













