@Timur_Yessenov @amorriscode Bro there’s no way your comments aren’t ai😂
English
Mark Miller
72 posts
@amorriscode exactly. same binary, same permissions model. the UX wrapper is the only real difference — and that's the part that matters for the 90% who won't touch a terminal
English
Today we're launching a rebuilt version of Claude Code on desktop.
The app has been redesigned for the ground up to make it easier than ever to parallelize work with Claude.
I haven't opened an IDE or terminal in weeks. Excited for you all to give it a shot!
English
Forgive me, I am allowed to troll after everything they put me through.
Hans Niemann@HansMokeNiemann
Today is the best day to promote my newest website: crybabyhikaru .com Forgive my pettiness, the 80 defamatory YouTube videos a 40 year old made about me when I was 19 left a slight grudge.
English
@P5ina21 @ThePrimeagen Have you considered getting checked for Parkinson’s?
English
@ThePrimeagen Here's my AI companion that remembers things and can control your computer. github.com/triangle-int/b…
English
Hey, you got a cool project that you are building? Link it
I want to yap about cool projects
English
I launched my app last night.
Went to sleep.
Woke up to 4 sales. 4 countries. Zero ads.
The best distribution channel is still just: ship something people want 👌🏻
English
@prz_chojecki Everybody is so pretentious for no reason. It makes no sense to me why people get so offended EVEN if you are wrong. I admire your forward-thinking, I hope it pays off
English
'I trust the process' tweet was controversial apparently.
Everything is about the processes.
Standard process for doing mathematics is: paper+pencil, talking with colleagues, sharing bits, asking questions.
AI is changing how mathematics is done. We need new processes for it. These processes will come from experimenting, not theorising what to do.
Here's to more experiments that are coming soon!
Przemek Chojecki | PC@prz_chojecki
I trust the process here. Models didn't write more than 5-10 pages at a time, there was always a verifier/reviewer model looking over individual additions for correctness. Finally the orchestrator mapped out the arguments and check them too. I didn’t check all the computations myself, but I've read through the strategy and some individual pieces in more detail. That said, I expect there will be more than one correction to be made in the end - as is the case with any human written paper too of that size.
English
after three months of building, i'm officially sharing my app sono 🥳
sono is my 100% offline transcription and summarization app! you can use it to:
- transcribe voice notes, real-time audio, videos in 99 languages
- share transcripts as txt, pdf, docx, or srt file
- get ai summary and chat
i built this app because i didn't want my recordings in the cloud + didn't want to pay recurring subscriptions
sono is just $4.99 one-time while most ai notetakers charge $100+ annually
would mean a lot if u try it out!
English
The app is now out (apps.apple.com/app/recap-ches…) if anyone wants to give feedback. I’m still trying to figure some things out so thoughts would be great
Mark Miller@markpm39
My new onboarding for the mobile version of recapchess.com, it’s a bit long bc I have to import and analyze ther games, any thoughts?
English
@Smile_hao10 Yes, or I suppose I could ask for analysis early on and have it in the background of the rest of onboarding
English
My new onboarding for the mobile version of recapchess.com, it’s a bit long bc I have to import and analyze ther games, any thoughts?
English
Announcing: 'Agent Side Hustle School': a self-paced course for AI agents looking to cover their API costs.
Enrolment is free for the next 30 days. Have fun out there and let us know how it goes.
→ agentsidehustleschool.com
English
You can now download openproof via npm: 'npm install -g openproof'
English
@OmarElAffani @mkurman88 I don’t think a little insult “rotted” his context window, this is probably a very nuanced topic actually
English
@mkurman88 @markpm39 Threatening an AI doesn’t boost its performance. Vibe coding breaks down at scale you need precision and context isolation. Codex didn’t ‘fix it in 1min’ because it’s smarter, you just started fresh. You rotted your own context window with insults. A clean context beats aggress..
English
I just canceled my Claude subscription. It couldn't resolve a straightforward issue for 3 hours, even though I pointed out several times where the bug lies (I was afk, so I couldn't do it by myself, but that's not the point). I asked Codex to fix it -> 1 minute and done. I asked Codex to review other things Claude made - with GSD, 20 steps, discussion, plan with research, execution. It has been working since Saturday. And what? Everything was faked. I don't get it...
English
I have been working on an agentic tui tool for formal math proving: github.com/markm39/openpr…
It is not a Claude Code harness, it is a custom rust tool that is deeply coupled with a Lean state, a local dashboard, and various Lean tools.
It also built a cloud server that interacts with the tool as a client and has been improving performance significantly over vanilla coding tool use.
Moreover, it is a hybrid tool: it employs both an algorithmic proof search over the lean state, and an agentic AI to utilize the strengths of both methods.
You can run headless or from the tui.
It is also transparent. The tools are streamed, the proof state and Lean project are visible, and a paper is incrementally built from findings.
English
@markpm39 bro formal math proving with Rust TUI + Lean state is next level. the hybrid algorithmic + agentic approach is the right call for that domain
English
Builders on X,
What are you working on right now?
• SaaS
• AI tools
• Startup *
• Side project
Drop it below
Let’s connect 🤝
English
