Amjad Masad

And like a lot of people just don't struggle with the fact that there's more to life than just like mechanistic physics of it all.

So when you start from this premise...

And I can talk about all the different, I can like have a technical debate about AI as well.

But just like at rock bottom, I feel like there's something special about humans.

There's something about the mystery of the world that we haven't really figured out.

I think there is a sense in which intentions matter and you can...

you can get into a self fulfilling theories of the world.

Right.

Uh, and we've seen that throughout, throughout history, uh, predicting doom and bringing about crisis, you know, happens a lot.

Um, and so I think it is important to be optimistic.

And I think if you have a, like a doomer mindset and, um,

it is dangerous for that reason, for sure.

And so I do think like people who are naturally depressed and like not very optimistic do kind of gravitate towards like a more doomer mindset.

So think about how machine learning models are trained.

Right now, machine learning models are trained by consuming super large corpus of content, essentially all of content on the Internet.

and then simulating, creating algorithms, so they would learn by creating internal algorithms, essentially that's a good way to understand machine learning, to simulate how a human might respond to a query.

You can go really far with that.

It's amazing that it works.

But if you give...

it queried on AI for something it doesn't have an algorithm for, it will fail at.