AI-related companies lost $190B in stock market value | Hacker News Reader

AI-related companies lost $190B in stock market value

futurism.com

35 pointsHoasi17 comments

When, over what period? This thing doesn't say, and isn't even dated itself even if it would say 'on opening this morning' or something.

It links to the Reuters report which contains this information. https://www.reuters.com/technology/ai-companies-lose-190-bil...

Honestly the Reuters report itself has more content than this post.

It certainly feels like AI is a new tech bubble. Lots of AI services just seem to be thin wrappers over OpenAI’s offerings.

I think the usage of the term "bubble" could use a little sharpening. In this context, for example, you could also use "gold rush" for an entirely different (yet similarly vague) effect.

Bubble implies an overall deflation. Gold rush implies a few people will get rich--especially those selling tools--and the rest will go belly-up. The latter seems more plausible in general. And, honestly, gold rush probably applies better to a lot of things in tech than bubble. (Though dot-com was probably more of a bubble even if some companies came out the other side.)

I wouldn't necessarily call it a bubble. There has been a breakthrough in AI recently, and now we see the new companies following that. Of course, 95% or so will fail, and the invested money is gone, but that happens basically with every new technology shift. From my point of view, what is needed for a real investment bubble is that the investors are not aware that there is a high failure rate of AI startups, but I don't think that is the case here.

For comparison, in case of the crypto bubble, there's really a huge community that thinks BTC or other cryptos will go "to the moon" and they seem to be immune to common investing fundamentals.

until you can run OpenAI’s offerings on your laptop.

https://www.semianalysis.com/p/google-we-have-no-moat-and-ne...

Which is itself just a titanic wrapping on a gigantic effort, to convert all human knowledge into a spreadsheet - and becomes instantly tained by echolalia, as the AI output poisons itself.

A bubble implies there's little or no actual value inside. I disagree completely.

As a senior software engineer that has fully integrated GPT-4 into my dev process, I can assure you there is definitely value here. I would say these LLMs (or where we go from here) are going to be at least as transformative as the smartphone.

Google missed, their stock tanked. Meanwhile MS is up 2% today and AMD is only down 0.3%.

Overall, although there is certainly AI hype, this article is rather meaningless.

The subtitle of the article is "HAVE WE ALREADY HIT PEAK AI?"

(I keep the caps because they had it.)

If we're talking about the peak of inflated expectations before we get to the Trough of Disillusionment, I think we might be close. I think the article over-indexes on a small set of data points but at the same time, I also don't hear the conflation of LLM and AGI nearly as much in the last month.

“When the capital development of a country becomes a by-product of the activities of a casino, the job is likely to be ill-done”

― John Maynard Keynes

MSFT up 60% in 1Y.

"Microsoft and Alphabet stocks have hit record highs following a year of AI hype. "

I believe what we are seeing here is simply that investors notice that Microsoft and Google are in such a powerful negotiation position that any value possibly generated by AI in the future will likely be captured by Msft/Goog and not by the actual AI startup inventing the tech.

I mean good luck rolling out any kind of enterprise / desktop app without Microsoft's blessing.

And is it even possible to build a useful AI assistant without access to the user's data, held hostage within Office 365 or Google Worksuite? You'll already be fully dependent on API access controlled by those monopolists on your launch day. And why wouldn't they charge you something like 90% of your revenue in access fees?

when AI winter came, how they gonna repurpose these dedicated hardwares?

Reply on news.ycombinator.com