At the moment, I have a ridiculous outlook on the industry's second decade following the founding of OpenAI ( 2025-2034 ). See, I've been living the past two decades with an understanding ( call that also ridiculous, if you will ) that people are much simpler than they believe themselves to be.
And so, I've watched the current trend in AI take off, with much glee - at least, now, we're out of the winter decades. However, the models remain bloated, because we still maintain a bloated understanding of human minds. At some point in the future, this bloat will flocculate, and communal comprehension of what a mind is will be simplified considerably from both, the numinous romantic model, and billion wheels of fortune model. In fact, I think we'll finally end up with algorithms which are different from the currently popular ones, running on highly supervised training data of several million data, tops. If such smaller systems aren't "perfectly human", at least they'll be pretty damn close. ( At which point, we may or may not be looking at a market crash such as the world hasn't seen in a while. )
Most people I've met are incredible simpletons. I don't see why a machine should be more complicated than that. The machines which are will ultimately be smarter, thank heavens.
Call me ridiculous. I've generally found myself to be such a thing.
-
AGI : what people get right - productivity rises. What people get wrong - everyone gets richer. Yes, in the long-run. In the short-run, we have to do this rearrangement of commerce because there will be no moat on talent. Do you understand what happens to businesses, when there is no moat on talent?
No comments :
Post a Comment