Deconstructing Rationality - Part 2
By Leo Gura - December 28, 2025 | 1 Comments
A serious philosophical exploration of the limits of rationality. All of the problems with the rationalist paradigm and how to transcend it to unlock higher levels of intelligence.
Saten says: In the video you say that AI models are intelligent in a fuzzy and intuitive way. I am an AI engineer myself and I know that those models are just numbers and computation, yet, the bigger they are the smarter they get (usually).
So that makes me wonder that its just a matter of more atoms and more computation.
Now AI beats humans with creative thinking, finding solutions to hardest problems. Recently in my team we had a very sneaky bug and AI agent was able to find it an explain it to us. It was so hidden and complex that we needed some work to comprehend it.
With the examples like “apples are healthy” we can say its “statistically true”
Like if we feed the AI all of the human experiences of entities eating any kind of apples it would be able to extract some deterministic rules about when is that statement about the apples true or no.