Everyone told me: “Don’t get into robotics.”
“It’s too hard, just go build a B2B SaaS.”
But you know me.
I only tried harder.
The next big leap is physical AI: AI models that interact with the real world.
But today, building AI for robots is almost impossible unless you're
The missing piece for the ChatGPT moment of robotics is finally here
Today, we are releasing phosphobot, an open-source app to control robots intuitively and train the most powerful AI models for the cheapest robots possible
-> phospho-app/phosphobot
🧵 What we are building:
Just told a robot what to do in plain English.
Launching today: Control robots with text prompts.
No coding. No setup. No expensive GPUs. No ML PhD needed.
Launch app, input prompt, BOOM—AI controls your robot.
Need more? Record examples, we fine-tune it.
Try now! Link below
I used to spend my nights fixing broken robotics datasets.
So we fixed the @LeRobotHF dataset format.
This is available today, in the open-source version of phospho. Still is 100% compatible with LeRobot.
🧵Why and how we did it:
I'm banned from attending this weekend’s robotics hackathon in Paris.
Over the years, I have participated in many hackathons. Sometimes won, sometimes didn't. Sometimes the project we shipped hit 1.4K Github stars. But loved it every time!
Don’t worry, I didn't do something
Europe rarely celebrates this kind of execution.
But we should.
Because Phospho is proof that robotics doesn’t have to be slow, expensive, or closed.
And that France is building… not just talking.
Thanks for having me, @plbiojout and @nicolasoulianov!
🍊YC interview simulator using OpenAI real-time API, built from my YC founder experience (minus @ndessaigne’s French accent)
- Famous 10-min format
- Grills you
- Spots pitch weaknesses
- Features @garrytan's avatar!
Know a YC applicant? Tag them below
Phospho développe des "cerveaux pour robots"
"Notre but, c'est de mettre des robots dans la vie des gens"
💬 @plbiojout , cofondateur de @phospho
🎙️ @Fsorel
MIT: Using ChatGPT might make you dumber.
MIT wired 54 students to EEG caps and watched their brains melt the moment ChatGPT started typing for them. The results are terrifying:
1. Neural dimming.
Unaided writers lit up the whole cortex; LLM users showed a flat-line in alpha &