Sep 23, 2026 · 55 min · 13 segments
Category Theory Meets Agentic AI You can now become a channel member on Youtube: [https://www.youtube.com/@aboutlogic](mailto:https://www.youtube.com/@aboutlogic) Your support helps us keep these…
Neil GhaniGuest
Thorsten AltenkirchHostDeniz SarikayaHost
And it just dawned on me one day that, well, actually, if you think about AI agents, and that's what the company I co-founded, which is called Codemi, does, and hopefully you'll allow me a few minutes to plug it at some point.

Um, we do agentic AI, and, um, well, what do AI agents do? I mean, they need to respond to prompts.

And then the really amazing thing is all the stuff we've done, Thorsten and I did, and then people like David Spivack have taken over, um, taken up on polynomial functors.

All this huge amount of work that's been done is totally applicable to agentic AI.

And the really nice thing is that, you know, 99% of the world is using a sort of untyped version of agentic AI.

Now, all of your listeners, being into logic, will understand the value of types.

But for reasons I don't know, the rest of the world has chosen to ditch, to avoid, to go around this most amazingly useful and simple technology, and our job at Codemi is to replace current untyped agentic AI with typed agentic AI, and containers are our technological scaffolding for doing that.
Maybe very naively, what's, like, the, the main, without trying to do advertisement, but what are the main benefits? Is it, like, security? Because, I mean, a lot of jailbreaking works with-

Well, I think the best way to explain this is just to go back to the very, very, very basic fundamentals.

And then types came along, and they were amazingly useful because you could find errors before they ran and caused problems.

And it just dawned on me one day that, well, actually, if you think about AI agents, and that's what the company I co-founded, which is called Codemi, does, and hopefully you'll allow me a few minutes to plug it at some point.

Um, we do agentic AI, and, um, well, what do AI agents do? I mean, they need to respond to prompts.

And then the really amazing thing is all the stuff we've done, Thorsten and I did, and then people like David Spivack have taken over, um, taken up on polynomial functors.

All this huge amount of work that's been done is totally applicable to agentic AI.

And the really nice thing is that, you know, 99% of the world is using a sort of untyped version of agentic AI.

Now, all of your listeners, being into logic, will understand the value of types.

But for reasons I don't know, the rest of the world has chosen to ditch, to avoid, to go around this most amazingly useful and simple technology, and our job at Codemi is to replace current untyped agentic AI with typed agentic AI, and containers are our technological scaffolding for doing that.
Maybe very naively, what's, like, the, the main, without trying to do advertisement, but what are the main benefits? Is it, like, security? Because, I mean, a lot of jailbreaking works with-

Well, I think the best way to explain this is just to go back to the very, very, very basic fundamentals.

And then types came along, and they were amazingly useful because you could find errors before they ran and caused problems.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.