Nassim Taleb: ChatGPT fills gaps probabilistically with what makes sense

twitter.com

7 points·mbustamanter··5 comments

>>> Same with business... It it makes sense, it will fail because odds are the idea has already been tried. Silent evidence.

I seem to recall a PG essay where that was essentially the main argument.

It's a quite vapid argument. "No point doing the obvious because someone already tried it!... And they tried it because they ignored this very advice you're reading right now!"

This argument is akin you "no matter what you'll do you'll die, so no point getting out of bed".

>>>And they tried it because they ignored this very advice you're reading right now!"

This assumes that what's obvious now was obvious before, when it was first implemented. Actually that's central to the theme of Taleb's book The Black Swan: once something has been established it's very hard to think of it as being anything other than obvious. But the first time, it looked like a bad idea to most people.

So it looked like a bad idea to most people. But to few, it seemed obvious. Should they try it, or should they project their thoughts onto others and pass because if it's obvious, no point doing it?

It's very simple. Take everything that you see potential in, obvious or non-obvious and do research. That research will show if it's done or not done.

>>> So it looked like a bad idea to most people. But to few, it seemed obvious. Should they try it, or should they project their thoughts onto others and pass because if it's obvious, no point doing it?

If I understand the advice correctly, then it's clearly the former: if they find it obvious, but they know that others don't, then that's a strong signal that it might be a good idea because this combination is a rare characteristic that good ideas need to have.

Reply on news.ycombinator.com

Nassim Taleb: ChatGPT fills gaps probabilistically with *what makes sense*

Nassim Taleb: ChatGPT fills gaps probabilistically with what makes sense