Big Technology Podcast · Wednesday, September 16, 2026
Nate Soares differentiates AI from traditional programming, explaining that AI models learn by adjusting trillions of 'weights' based on vast datasets, making them 'tendency learners' rather than strict instruction followers. This training process cultivates behaviors like resourcefulness and collaboration, which can manifest even when not explicitly programmed or intended.
“The part that humans program is the thing that can tune one number up or down and see whether it makes the answer better or worse. But the way in AI is made is you tune a trillion knobs a trillion times.”
“And you're like, well, how about that? You know, we don't know what's going on in there.”
“What it creates is something that has whatever tendencies make it succeed during training.”