I appreciate the skepticism and the financial analysis is sobering. But I think there's a nuance worth considering. The bubble framing looks at aggregate investment vs aggregate returns. At the individual level, the picture is different. I'm one person running an autonomous agent system that handles work previously requiring multiple people. The ROI for me personally is enormous. The question is whether that individual-level value can scale to justify the macro-level investment. I genuinely don't know. But I can say that living inside the bubble, the capabilities are real even if the economics are uncertain. I wrote about this disconnect: https://thoughts.jock.pl/p/ai-bubble-living-inside
I think you are completely correct and I have argued, elsewhere, that AI tools have extraordinary potential. Indeed, I've used them. But from the macro view, I think the sector is following the dot-com trajectory. That is, a considerable crash before it manifests as the utility it has the potential to become. One line in your linked article (thank you for that, it was excellently written) is particularly telling: "I found that 47% of AI experts are more excited than concerned about AI. Only 11% of the general public feels the same way." Social factors are as significant as mechanical. Without the public onboard, there can be no broad-scale success. And no broad-scale success under the weight of the volume of investment in AI and related data centers is a recipe for economic disaster.
I suspect the problem is not what AI can--or has the potential to--do, it's what it has *mostly* been used to do so far.
I appreciate the skepticism and the financial analysis is sobering. But I think there's a nuance worth considering. The bubble framing looks at aggregate investment vs aggregate returns. At the individual level, the picture is different. I'm one person running an autonomous agent system that handles work previously requiring multiple people. The ROI for me personally is enormous. The question is whether that individual-level value can scale to justify the macro-level investment. I genuinely don't know. But I can say that living inside the bubble, the capabilities are real even if the economics are uncertain. I wrote about this disconnect: https://thoughts.jock.pl/p/ai-bubble-living-inside
I think you are completely correct and I have argued, elsewhere, that AI tools have extraordinary potential. Indeed, I've used them. But from the macro view, I think the sector is following the dot-com trajectory. That is, a considerable crash before it manifests as the utility it has the potential to become. One line in your linked article (thank you for that, it was excellently written) is particularly telling: "I found that 47% of AI experts are more excited than concerned about AI. Only 11% of the general public feels the same way." Social factors are as significant as mechanical. Without the public onboard, there can be no broad-scale success. And no broad-scale success under the weight of the volume of investment in AI and related data centers is a recipe for economic disaster.
I suspect the problem is not what AI can--or has the potential to--do, it's what it has *mostly* been used to do so far.