4 ms·

The problem is that the bubble people are so unimaginative, similar to Krugman, that those who have any inkling of an imagination can literally feel like vision

by ivape 11mo ago

The problem is that the bubble people are so unimaginative, similar to Krugman, that those who have any inkling of an imagination can literally feel like visionaries compared to them. I know I’m describing Dunning-Krueger, but so be it, the bubble people are very very wrong. It’s like, man, they really are unable to imagine a very real future.

teaearlgraycold 11mo ago
Almost everyone I hear calling our AI hype machine a bubble aren't claiming AI is a short term fluke. They're saying the marketing doesn't match the reality. The companies don't have the revenue they need. The model performance is hitting the top of the S curve. Essentially, this is the first big wave - but it'll be a while before the sea level rises permanently.
- bdangubic 11mo ago
  > marketing doesn't match the reality. true for every marketing ever
  - an0malous 11mo ago
    It’s not just a marketing stunt, it’s a trillion dollar grift that VCs are going to try to dump off onto the public markets when the reality doesn’t catch up to the hype fast enough
bccdee 11mo ago
I find the argument for the bubble to be extremely straightforward. Currently, investment into AI exceeds the dot-com bubble by a factor of 17. Even in the dot-com era, the early internet was already changing media and commerce in fundamental ways. November is the three-year anniversary of ChatGPT. How much economic value are they actually creating? How many people are purchasing AI-generated goods? How much are people paying for AI-provided services? The value created here would have to exceed what the internet was generating in 2000 by a factor of 17 (which seems excessive to me) to even reach parity with the dot-com bubble. "But think where it'll be in 5 years"—sure, and let's extrapolate that based on where it is now compared to where it was 3 years ago. New models present diminishing returns. 3.5 was groudbreaking; 4 was a big step forward; 5 is incremental. I won't deny that LLMs are useful, and they are certainly much more productized now than they were 3 years ago. But the magnitude of our collective investment in AI requires that a huge watershed moment be just around the corner, and that makes no sense. The watershed moment was 3 years ago. The first LLMs created a huge amount of potential. Now we're realizing those gains, and we're seeing some real value, but things are also tapering off. Surely we will have another big breakthrough some day—a further era of AI which brings us closer to something like AGI—but there's just no reason to assume AGI will crop up in 2027, and nothing less that that can produce the ROI that such enormous valuations will eventually, inexorably, demand.
- tim333 11mo ago
  That "factor of 17" comes from an interest rate model that is unrelated to AI.
  - lucaslazarus 11mo ago
    This is not true. Obviously the underlying effect is real but not nearly to this scale—for instance, neither the CPI nor the S&P500 are even remotely close to 17x higher than they were at the turn of the millennium.
    - tim333 11mo ago
      The source is a report written by Julien Garran based on the difference between actual interest rates and an idea of what they should be called the Wicksell spread. There's a summary here https://www.marketwatch.com/story/the-ai-bubble-is-17-times-the-size-of-the-dot-com-frenzy-this-analyst-argues-046e7c5c https://www.marketwatch.com/story/the-ai-bubble-is-17-times-... He figured there was a credit bubble like that around the time of the dot com bubble and now but the calculation if purely based on interest rates and the money can go into any assets - property, stocks, crypto etc. It's not AI specific. He explains it here https://youtu.be/uz2EqmqNNlE https://youtu.be/uz2EqmqNNlE The Wicksell spread seems to have come from Wicksell's proposed 'natural rate of interest' detailed in his 1898 book https://en.wikipedia.org/wiki/Knut_Wicksell#Interest_and_Prices,_1898 https://en.wikipedia.org/wiki/Knut_Wicksell#Interest_and_Pri...
      - lucaslazarus 11mo ago
        I see, thank you!
- lucaslazarus 11mo ago
  I don’t get why people find it so hard to understand that a technology can be value-additive and still be in a position of massive overinvestment. Every generation of Californians seeks to relive the 1848 gold rush, spending millions excavating rivulets for mere ounces of (very real!) gold.
techblueberry 11mo ago
It’s a weird comparison since internet in the dial-up age was a bubble, are you saying the hype machine for AI is in fact smaller than the internet? Are you implying that AI will in fact grow that much more slowly and sustainably than the internet, despite trillions of investment? Do you think Sam Altman, Jeff Bezos, and Mark Zuckerberg are all wrong saying that we’re in a bubble? Do they “lack imagination?” Also? What do I need imagination for, isn’t that what AI does now?
- timeinput 11mo ago
  That’s a sharp and layered question, and I think you’re cutting right to the heart of the current tension around AI.
- Libidinalecon 11mo ago
  Not only was it a bubble but we actually had the internet. We didn't have to use our imaginations and build for the technology yet to be invented and we still had a massive damaging bubble. Even beyond that, the Soviet Union had just collapsed. The US had a balanced budget and 10X less debt. Globalization was just getting started. China was nothing compared to what it was today. It was the absolute most stable time of my life. Basically the opposite of 2025. "This time things are different because AI blah blah blah". When you boil it down, all bubbles are based around the idea of using your imagination to image how "this time it is different". The people who don't think this is a bubble are not seeing that their imagination is a bug and not a feature here.

by ivape 11mo ago

teaearlgraycold 11mo ago

Almost everyone I hear calling our AI hype machine a bubble aren't claiming AI is a short term fluke. They're saying the marketing doesn't match the reality. The companies don't have the revenue they need. The model performance is hitting the top of the S curve. Essentially, this is the first big wave - but it'll be a while before the sea level rises permanently.

bdangubic 11mo ago
> marketing doesn't match the reality. true for every marketing ever
- an0malous 11mo ago
  It’s not just a marketing stunt, it’s a trillion dollar grift that VCs are going to try to dump off onto the public markets when the reality doesn’t catch up to the hype fast enough

bccdee 11mo ago

I find the argument for the bubble to be extremely straightforward. Currently, investment into AI exceeds the dot-com bubble by a factor of 17. Even in the dot-com era, the early internet was already changing media and commerce in fundamental ways. November is the three-year anniversary of ChatGPT. How much economic value are they actually creating? How many people are purchasing AI-generated goods? How much are people paying for AI-provided services? The value created here would have to exceed what the internet was generating in 2000 by a factor of 17 (which seems excessive to me) to even reach parity with the dot-com bubble. "But think where it'll be in 5 years"—sure, and let's extrapolate that based on where it is now compared to where it was 3 years ago. New models present diminishing returns. 3.5 was groudbreaking; 4 was a big step forward; 5 is incremental. I won't deny that LLMs are useful, and they are certainly much more productized now than they were 3 years ago. But the magnitude of our collective investment in AI requires that a huge watershed moment be just around the corner, and that makes no sense. The watershed moment was 3 years ago. The first LLMs created a huge amount of potential. Now we're realizing those gains, and we're seeing some real value, but things are also tapering off. Surely we will have another big breakthrough some day—a further era of AI which brings us closer to something like AGI—but there's just no reason to assume AGI will crop up in 2027, and nothing less that that can produce the ROI that such enormous valuations will eventually, inexorably, demand.

tim333 11mo ago
That "factor of 17" comes from an interest rate model that is unrelated to AI.
- lucaslazarus 11mo ago
  This is not true. Obviously the underlying effect is real but not nearly to this scale—for instance, neither the CPI nor the S&P500 are even remotely close to 17x higher than they were at the turn of the millennium.
  - tim333 11mo ago
    The source is a report written by Julien Garran based on the difference between actual interest rates and an idea of what they should be called the Wicksell spread. There's a summary here https://www.marketwatch.com/story/the-ai-bubble-is-17-times-the-size-of-the-dot-com-frenzy-this-analyst-argues-046e7c5c https://www.marketwatch.com/story/the-ai-bubble-is-17-times-... He figured there was a credit bubble like that around the time of the dot com bubble and now but the calculation if purely based on interest rates and the money can go into any assets - property, stocks, crypto etc. It's not AI specific. He explains it here https://youtu.be/uz2EqmqNNlE https://youtu.be/uz2EqmqNNlE The Wicksell spread seems to have come from Wicksell's proposed 'natural rate of interest' detailed in his 1898 book https://en.wikipedia.org/wiki/Knut_Wicksell#Interest_and_Prices,_1898 https://en.wikipedia.org/wiki/Knut_Wicksell#Interest_and_Pri...
    - lucaslazarus 11mo ago
      I see, thank you!
lucaslazarus 11mo ago
I don’t get why people find it so hard to understand that a technology can be value-additive and still be in a position of massive overinvestment. Every generation of Californians seeks to relive the 1848 gold rush, spending millions excavating rivulets for mere ounces of (very real!) gold.

techblueberry 11mo ago

It’s a weird comparison since internet in the dial-up age was a bubble, are you saying the hype machine for AI is in fact smaller than the internet? Are you implying that AI will in fact grow that much more slowly and sustainably than the internet, despite trillions of investment? Do you think Sam Altman, Jeff Bezos, and Mark Zuckerberg are all wrong saying that we’re in a bubble? Do they “lack imagination?” Also? What do I need imagination for, isn’t that what AI does now?

timeinput 11mo ago
That’s a sharp and layered question, and I think you’re cutting right to the heart of the current tension around AI.
Libidinalecon 11mo ago
Not only was it a bubble but we actually had the internet. We didn't have to use our imaginations and build for the technology yet to be invented and we still had a massive damaging bubble. Even beyond that, the Soviet Union had just collapsed. The US had a balanced budget and 10X less debt. Globalization was just getting started. China was nothing compared to what it was today. It was the absolute most stable time of my life. Basically the opposite of 2025. "This time things are different because AI blah blah blah". When you boil it down, all bubbles are based around the idea of using your imagination to image how "this time it is different". The people who don't think this is a bubble are not seeing that their imagination is a bug and not a feature here.