Ondrej
@HeroStudioAI
Followers
226
Following
356
Media
33
Statuses
138
Building https://t.co/dZ6cV3Vfge in public 🚧 Terminal-first AI screen recorder for developers CLI → Record → AI Edit → Ship cinematic demos
London
Joined July 2025
RT @HeroStudioAI: The reason I prefer Reddit much more is that there actually are some people and your posts are visible. It wasn’t obvious….
0
1
0
The reason I prefer Reddit much more is that there actually are some people and your posts are visible. It wasn’t obvious until they recently started showing the insights. And X? It’s totally dead social network. Looks like everyone left after @elonmusk took over.
0
1
0
This channel is “build in public” but I very rarely see someone sharing something about building. What you are working on? What tech do you use? What’s unique? What’s your expertise? What’s your progress.
1
0
3
If you use AI to help with code, please do tests and do a lot. You get it almost for free so why would someone choose not to do tests?.
1
0
6
OBS is interesting piece of software, not shiny, a bit oldschool but unbelievable powerful and reliable ✨.
0
0
0
work heavily in progress but finally getting into composition stage. You can see the cursor is off. The subtitles not finished. The camera glass needs to be more "glassy" . still lot to do. But most important this all is recorded and processed in less than 10 seconds including
0
0
0
RT @HeroStudioAI: This is HUGE. Google DeepMind released URL Context. You can now extract content from any webpage, PDF, or image just by….
0
1
0
RT @HeroStudioAI: One of the effects I was working on today. I call it "frozen glass" (as a contrast to apple's liquid glass 😃). https://t.….
0
1
0
RT @HeroStudioAI: Among all the vibe coders, are there any professional software engineers too?.
0
1
0
Among all the vibe coders, are there any professional software engineers too?.
0
1
2
One of the effects I was working on today. I call it "frozen glass" (as a contrast to apple's liquid glass 😃).
0
1
0
You don’t need a GPU to run 100B+ models anymore. Microsoft just open-sourced bitnet.cpp, a 1-bit inference framework that makes massive LLMs run efficiently on CPUs:. 👉 6× faster inference.👉 82% lower energy use.👉 Supports Llama 3, Falcon 3, and BitNet models. Massive.
0
0
0
This is HUGE. Google DeepMind released URL Context. You can now extract content from any webpage, PDF, or image just by pasting a URL. It pulls live data from up to 20 links per request! 🤯. No setup needed. Just pass the links in your prompt. Here are a few powerful use.
0
1
3
Switch filter from TRENDING to NEW and it suddenly gives you much more interesting posts. X algo is getting gradually more 💩.
2
0
4
RT @HeroStudioAI: 👋 Hey builders!. I’m working on a terminal-first video editor using:. 🦀 Rust · 🐍 Python · 💻 TypeScript .🧠 Computer Visio….
0
1
0