Comment by addisonj

Comment by addisonj 8 hours ago

I will repeat my comment from 70 days ago:

> I was discussing with a friend that my biggest concern with AI right now is not that it isn't capable of doing things... but that we switched from research/academic mode to full value extraction so fast that we are way out over our skis in terms of what is being promised, which, in the realm of exciting new field of academic research is pretty low-stakes all things considered... to being terrifying when we bet policy and economics on it.

That isn't overly prescient or anything... it feels like the alarm bells started a while ago... but wow the absolute "all in" of the bet is really starting to feel like there is no backup. With the cessation of EVs tax credits, the slowdown in infra spending, healthcare subsidies, etc, the portfolio of investment feels much less diverse...

Especially compared to China, which has bets in so many verticals, battery tech, EVs, solar, then of course all the AI/chips/fabs. That isn't to say I don't think there are huge risks for China... but geez does it feel like the setup for a big shift in economic power especially with change in US foreign policy.

matthewaveryusa 7 hours ago

I'll offer two counter-points. Weak but worth mentioning. wrt China there's no value to extract by on-shoring manufacturing -- many verticals are simply uninvestable in the US because of labor costs and the gap of cost to manufacture is so large it's not even worth considering. I think there's a level of introspection the US needs to contend with, but that ship has sailed. We should be forward looking in what we can do outside of manufacturing.

For AI, the pivot to profitability was indeed quick, but I don't think it's as bad as you may think. We're building the software infrastructure to accomodate LLMs into our work streams which makes everyone more efficient and productive. As foundational models progress, the infrastructure will reap the benefits a-la moore's law.

I acknowledge that this is a bullish thesis but I'll tell you why I'm bullish: I'm basically a high-tech ludite -- the last piece of technology I adopted was google in 1996. I converted from vim to vscode + copilot (and now cursor.) because of LLMs -- that's how transformative this technology is.

Reply View 5 replies

vivalahn 2 minutes ago

> We should be forward looking in what we can do outside of manufacturing.
For example?

Reply View | 0 replies
hakfoo 22 minutes ago

I think an interesting way to measure the value is to argue "what would we do without it?"
If we removed "modern search" (Google) and had to go back to say 1995-era AltaVista search performance, we'd probably see major productivity drops across huge parts of the economy, and significant business failures.
If we removed the LLMs, developers would go back to Less Spicy Autocomplete and it might take a few hours longer to deliver some projects. Trolls might have to hand-photoshop Joe Biden's face onto an opossum's body like their forefathers did. But the world would keep spinning.
It's not just that we've had 20 years more to grow accustomed to Google than LLMs, it's that having a low-confidence answer or an excessively florid summary of a document are not really that useful.

Reply View | 0 replies
827a 7 hours ago

Another thing to note about China: while people love pointing to their public transit as an example of a country that's done so much right, their (over)investment in this domain has led to a concerning explosion of local government debt obligations which isn't usually well-represented in their overall debt to GDP ratios many people quote. I only state that to state that things are not all the propaganda suggests it might be in China. The big question everyone is asking is, what happens after Xi. Even the most educated experts on the matter do not have an answer.
I, too, don't understand the OP's point of quickly pivoting to value extraction. Every technology we've ever invented was immediately followed by capitalists asking "how can I use this to make more money". LLMs are an extremely valuable technology. I'm not going to sit here and pretend that anyone can correctly guess exactly how much we should be investing into this right now in order to properly price how much value they'll be generating in five years. Except, its so critical to point out that the "data center capex" numbers everyone keeps quoting are, in a very real (and, sure, potentially scary) sense, quadruple-counting the same hundred-billion dollars. We're not actually spending $400B on new data centers; Oracle is spending $nnB on Nvidia, who is spending $nnB to invest in OpenAI, who is spending $nnB to invest in AMD, who Coreweave will also be spending $nnB with, who Nvidia has an $nnB investment in... and so forth. There's a ton of duplicate-accounting going on when people report these numbers.
It doesn't grab the same headlines, but I'm very strongly of the opinion that there will be more market corrections in the next 24 months, overall stock market growth will be pretty flat, and by the end of 2027 people will still be opining on whether OpenAI's $400B annual revenue justifies a trillion dollars in capex on new graphics cards. There's no catastrophic bubble burst. AGI is still only a few years away. But AI eats the world none-the-less.
[1] https://www.sciencedirect.com/science/article/abs/pii/S09275...

Reply View | 2 replies
- addisonj 6 hours ago
  
  My point is not that value extraction wouldn't happen, my point is simply that in addition to the value extraction we also made other huge shifts in economic policy that taken together really seem to put us on a path towards an "AGI or bust" situation in the future.
  Is that a bit hyperbolic? isn't this just the same as dotcom and housing bubbles before where we pivoted a bit too hard into a specific industry? maybe... but I also am not sure it would be wise to assume past results will indicate future returns with this one.
  
  Reply View | 1 reply
  
  hakfoo 9 minutes ago
  
  AI is appealing to the investors not because it solves human problems, but because it solves some of the problems of previous bubbles.
  When we wired the world for the Internet in the 1990s, or built railways across the continent in the 1800s, we eventually reached a point where even the starriest-eyed investors could see they've covered effectively the entire addressable market. Eventually AOL ran out of new customers no matter how many CDs they mailed out, or we had connected every city of more than 50 people with steel rail, and you could hear the music was slowing down.
  By dangling the AGI brass ring out there, they can keep justifying the expenditure past many points of diminishing returns, because the first thing we'll ask the Omnipotent AGI is how to earn the quadrillions spent back, with interest.
  It also has the benefit of being a high-churn business. The rails laid in 1880, or the fiber pulled in 2000, were usable for decades, but in the AI bubble, the models are obsolete in months and the GPUs in years. It generates huge looking commercial numbers just to tread water.
  
  Reply View | 0 replies

klooney 8 hours ago

> but geez does it feel like the setup for a big shift in economic power

It happened ten years ago, it's just that perceptions haven't changed yet.

Reply View 0 replies