Verifiable feedback loops accelerate AI progress
The guest argued that AI capabilities advance most rapidly and predictably in fields with objective, verifiable feedback loops like software engineering and materials science.
The argument
Anjane Midha explained that structured pipelines—such as unit testing in coding or robotic physical testing in materials science—allow models to train on factual data and avoid hallucinations. Conversely, progress remains slow and prone to errors in subjective domains like creative writing or therapy.
The thesis, stress-tested
✓ What validates it
- ✓Commercial viability of AI-discovered materials such as superconductors
- ✓Widespread adoption of fully autonomous AI software engineers passing enterprise unit tests
▸ Risks discussed
- ▸Subjective tasks remain highly prone to hallucinations
- ▸Models may struggle to generalize outside of highly structured environments
Hear it yourself
"But, you know, it'd be helpful to get some of your advice on how to do that. And I couldn't really come on board full time at the time with them because, I had to integrate my company to the acquirer. Sure. But I came on as their angel, and nights and weekends, I worked with them on the business plan and and who we should raise from. That, you know, that company is entropic. Dario and Tom and I started doing these weekly working sessions in early twenty twenty one. And, yeah, we I I assumed that, you know, if we went and talked to a bunch of venture capitalists on Sandal"
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE