Ramade: Machine Learning Models Cannot Learn Physics Probabilistically
Pratap Ramade · How AI Can Enable AMBITION| | Pratap Ramade on TBPN April 8th · Apr 10, 2025 · at 9:33
Arena co-founder Pratap Ramade explains why engineering AI requires hardcoded physical constraints rather than relying purely on probabilistic ML training.
0:00 / 0:19exact quote · 19.0s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“There are certain rules of the world that we've learned over time that need to be true. It's like gravity is 9.8 meters per second square. You can't probabilistically learn that by watching stuff fall in air and being like, yeah, my ML. No, no. I mean, there's just some like speed of light. Like you're going to encounter that shit. It doesn't matter if you're an ML model or like, you know, it's just real, right?”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Pratap Ramade
Ramade: Horizontal AI is not a winning game for small startups
“Look, I don't think for a small company playing in horizontal AI is really a winner's game”
Ramade: Hardware lab tooling has stagnated under 1980s incumbents
“The incumbent competitor set
Are three companies from the eighties.
You know, it's weirdly the underpinning layer of technology on which all of our software runs weirdly hasn't changed that much.
Developing the hardware has like stagnant
and it's kind of surpr…”
Ramade: Entire SaaS companies will become mere software features
“What we considered a SaaS company is now just going to be a feature for in the future.”
Ramade: AI capability for hands-on engineering tinkering remains disastrous
“The tinkering in garage problem is very unsolved. Like you look at AI's capability there. It's like, it's a disaster.”