Vasant Dhar

So, you know, I think we're close.

And, you know, we're there where AI can actually play a pretty significant role in sort of long-term systematic investing.

So, you know, the traditional view of machine learning is supervised learning.

You have examples that you learn from.

There just weren't sufficient examples or training data to learn from, right?

You know, Damodaran is sort of exceptional in terms of the number of reports he's written,

you know, which go into sort of the low thousands maybe at this point.

But that isn't sufficient data to really train an AI, like, you know, compared to how much data there was in the high frequency in the short-term space.

So there just wasn't enough training data.

What's different now, you know, and this is something I write in my book, is this emergence of general intelligence, where the machine knows something about everything, and the division between

common sense reasoning and expertise is broken down, right?

To me, that's like the big deal about AI that's made it a general purpose technology.

That was always the hindrance, you know, that we sort of drew this artificial boundary between expertise and common sense.

And with LLMs, that became available.

So now you can build on this sort of substrate of, you know, a machine that has common sense

And you don't need as much training data because it's ingested sort of the wisdom of humanity on the Internet.

And now it's a question of like, you know, tilting it in an appropriate direction, such as a value or a fundamentals approach like the modern does.

Right.

And that's what's made it feasible for, you know, reasoning about sort of long, longer term investment horizons.

So I've discussed the SpaceX valuation with him.