Replying to @jack
It’s impressive, but it feels like it’s trying too much to be human, gives me the creeps.
Replying to @Revolt3r
Maybe if there is enough interest, but also open for PR’s to add Win/Linux versions
Replying to @d4m1n
🫡 Let me know how it goes and if you like to add some feature or improvements.
Replying to @PovilasKorop
When using the same models you will get a good understanding of their capabilities and how they perform in your codebase, so you know what goes through quickly, what likely needs a lot back and forth and what is time consuming with QA.
Replying to @tannerlinsley
The agent harnesses are amazing, been building parts of that myself since the start of the year, replaced most of it with Tanstack AI now. Top tier stuff.
Replying to @OpenAIDevs
The Decision API is incredible it's classifying mutations and faults in cell mitosis for my cancer research tool. It's incredibly cheap to run.
xitter.cf/eelcodotdev/status/211…
Never expected to be working on actual Cancer research. I'm absolutely amazed what a massive unlock Opus 5.5 is, we should all be using it for the better of humanity.
This tool will save researches many hours every week recording cell devisions, detecting faults and mutations.
Built with @claude, running on @OpenAI decision API.
I'm using it to spin up claude and codex sandboxes on trevize.dev
Alright we're in action it seems now, you don't have to leave X anymore to do your work.
@bot what's up!
Replying to @d4m1n
The design was one shot, it's already better than many alarms in the app store right now (aside from the usual ai slop copy)
Claude now has IOS simulator support, it's really good.
Building a wake up alarm with generative audio. I'm tired of the same wake up sounds every single day.
Replying to @lightsilver323
Share a folder with your Grok @bot.
botdrive.app/





