this, and other generative language models really is getting me back into following any kind of technology. It’s mind blowing.

@manton I building Chat GPT functionality into ClipDish and last night I found myself just chatting away with my "sous-chef."

@manton I took an online Coursera class on machine learning, so I have at least a very superficial understanding of the concept. And yes, even at that primitive level, it’s already basically magic, as in, no one actually knows what’s going on inside the box. Wolfram’s penultimate blog post explains how it works, and his latest says the inner complexity maybe doesn’t matter, what matters is what you can do with it.

Is it because building an AI is very different from building a software?

As software developers, we are used to computers being deterministic machines. The statistical models backed by very large datasets can seem different, but it’s still just code under the hood.