Hacker News Reader

Comment by deafpolygon | Hacker News Reader

AI in its current state is a bubble. It's going to pop soon, and then we'll get back to business as usual.

I don’t disagree about a bubble. A bubble for promising technology is probably inevitable. But what is this about going back to “business as usual”? Now that GenAI exists it has fundamentally changed what tasks a computer is good at and which use cases are expensive vs. cheap. So there’s no going back. Even if all models stopped improving tomorrow, we’ll be adapting our products for 10 years to take advantage of this fundamental change in capabilities.

Now all that is a long way away from creating an AI god which is what the peak hypers are selling.

A few years ago saying that there will be soon AI that perfectly understands human natural language was viewed as unrealistic science fiction. Today the possibility of an AI god is regarded as unrealistic science fiction.

I keep watching old Star Trek TNG episodes and the way Data and the ship’s computer is portrayed seems to have been largely matched or even surpassed. And that was supposed to be 24th century tech that we caught up to in 35 years.

Todays models don't hold a candle to data, they are better at reading emotions yes but everything else they are horrible at in comparison.

Sci fi writers just assumed that emotions would be hard for computers but apparently it isn't that hard at all compared to rational thinking.

We're nowhere near Data, or ship's computer, level of AI.

Is this a joke?

Data has the ability to be accurate and store and index all the information, not present a lossy copy where he hallucinates.

Just to clarify a few things. First, there is no AI, only LLMs. Second, no LLM understands anything at all, never mind human language. Third, no LLM operates anywhere near to perfection at any task. It is still unrealistic science fiction.

Wishful thinking.

Just to clarify a few things. First, there is no intelligence just evolved primates triggering neurons with chemicals. Second, no human understands anything at all they're just dumb machines reacting to stimulus. Third, no human operates anywhere near perfection at any task.

I take it you mean that like the dotcom bubble? As in The Internet didn't go away, but the investors returned to a little more sensible thinking.

LLMs will continue to be used by increasingly more developers

I'd agree with you except that their product is insanely good compared to the rest of the world, and they're leading into a new space as well.

Sure the bubble will pop with the hundreds of nonsense startups going on at the moment, but OpenAI will be fine.

Soon is what timeline? It's still getting better; as long as that happens, no one is stopping.

What does this even mean, really?

It means that the return of investment for fancy email autocomplete is not that big

Please keep the cheeky reductionist takes to yourself

Seed rounds for Ai startups with nothing more than a napkin saying "do Ai stuff" will have valuations less than a billion.

Reply on news.ycombinator.com