Hacker News Reader

Comment by EA-3167 | Hacker News Reader

The cynic in me says this is about getting the money before the hype starts to fade and people begin to see that these "AI" are not leading to what people mean when they say AI.

View on HN

Kind of reminded me of how just before WeWork came crashing down it raised more money at a $47 billion valuation. Besides the fact that both companies have (or rather, had) lofty valuations, there are some other interesting parallels. There's the lack of any real moat. There's the insane amounts of cash they were spending, on leases in WeWork's case, and tech infra in the case of OpenAI. Both were lead by highly charismatic CEO's who were arguably their respective company's greatest asset, but also perhaps their greatest liability.

Kind of also reminded me of how Andreesen Horowitz gave Adam Neumann $350 million back in 2022 for some new real-estate venture. I mean, at this point, it seems very stupid to give him more money, and yet people who appear to be "in the know" are not only giving him more money, but absolutely insane amounts of money.

I bring up this latter point because it's one of those things that kind of makes you question what you understand. Why are otherwise smart and knowledgeable people doing something that seems stupid? Are you wrong? Or are they wrong? I mean, they know more than you do, so normally you would be wrong, but in this case what they're doing seems so stupid that you aren't so sure.

I feel similarly about AI. Not that I think AI is necessarily stupid. I for one am excited about what could come from applying AI to drug discovery. It just seems kind of crazy to create a machine that's really good at next token prediction and extrapolate from that all these crazy predictions. That this will lead to machines capable of understanding, reasoning, intelligence, and eventually, AGI.

Either the experts are wrong, or I'm wrong, and I'm normally inclined to believe in the experts, but in this case what they believe is so nutty that I really have to question things.

I think the key is that believing a consensus of experts is often a good default, but experts remain human, and human are corruptible for much less money than billions of dollars. For that sort of money anyone can convince themselves of pretty much anything, especially when the market is offering feedback in the form of, "You're brilliant, take our money!"

It's not AGI but it's still really impressive. It has the feel of the nascent personal computer industry or the internet pre-dotcom. Yes, there's hype but there's legitimate value propositions in play.

But there’s a difference between “impressive” and “100B impressive”, no? It’s not clear how big the market will be (depends on tech advances) and how much of it they can retain.

Yeah, there's a definite dotcom vibe in the pricing. It's also clear that there's too much money out there looking for a home.

Its impressive to me as well but does it really have a moat though to feel safe investing in?

In fact I think it’s more useful than AGI will be. IMO AGI will be a novelty, but generative AI has exceptional potential as a tool.

This is a wild comment

If you showed GPT-3.5 to Claude Shannon he would say it’s what they were trying to create

I took the poster to mean people in general, as in those who think of AI as defined by Hollywood, not the people who actually work in AI.

100%, but I wouldn’t characterize yourself as cynical

This pattern just happens over and over in VC/startup land. See IoT and cryptocurrency/NFT companies as previous example in previous tech hype cycles.

I am of the persuasion that secret capabilities and even conceptual roadmaps or stretch goals that they are showing to [them] are compelling to the point that they are practically building OpenAI an office at the US Mint.

I mean even the simplest of "what if" scenarios that even the HN crowd can think up can be pretty 'entangling'.

I thought of something after listening to Aza Rasking and Tristan Harris on JRE and their Summit talk...

I thought of what brought me along in tech, and I believe it is the "what ifs"

I wrote this down after listening to Aza:

Neuromancer and my trusty cyberpunk persona and fascination with computers got me into what I did my whole career. As so many of us 80's kids mantra'd as we read and dreamed "Wouldn't it be cool if!?"

And so we made it.. and others saw it, and said "wouldn't it be cool if?!" and made it more and better and cooler - until others saw it and we all got together and thought "Wouldn't it be cool if?!"

And we built it together, and competed for more and more cool-if's ...

And we didn't notice the entanglements. We just wanted the cool-if's.

But now, wouldn't it be cool if we could untangle the cool, and just be.*

---

https://www.youtube.com/watch?v=cB0_-qKbal4

(JRE Spotify talk is better than this one though)

Reply on news.ycombinator.com