[conartist6]

There can be a bubble of speculation around something that does have real value. Housing for example.

What is driving the market to assign such a huge value to models is that they can sell themselves as the solution to any and every problem, even if they aren't

[paulcole]

I don't understand what you mean by this. What specifically do you think is going to happen?

[TheNewsIsHere]

Not OP, but what typically happens: a large contraction.

Investors are already contracting the over-investment in AI because it’s so hard to turn an ROI. Satya Nadella was recently on the record as saying he doesn’t think AI has proven a sustainable ROI yet, and that it needs to in order to sustain the current levels of investment. As a matter of cause and effect he is of course right. But no one can predict the timelines and outcomes with exactitude.

Tomorrow Nvidia could be seized by a regulator and frozen, decimating the GPU market. SoftBank could decide they’re going all in on some new thing instead. (Incredible examples, of course.)

Personally I don’t think AI/LLMs/et al are going away. But the way that it’s getting bolted into -everything- is likely to eventually subside.

It’s been interesting to see companies rebrand around AI even when they don’t actually produce or integrate any of those technologies. That’s driving part of the overvaluation of AI at large.

AI and the collection of technologies that get called AI aren’t new. They’re just suddenly in focus and all the rage.

We’ve seen this cycle before in technology and we’ll see it again with something else.

[nyarlathotep_]

Seems we're starting to see it now:

https://www.datacenterdynamics.com/en/news/wells-fargo-amazo...

[lioeters]

The hype bubble bursts: market experiences a trough of disappointment, then recovers a slope of enlightenment toward a realistic valuation of the new technology.

[paulcole]

I don't understand what you mean by this. What specifically do you think is going to happen?