you are viewing a single comment's thread
view the rest of the comments
[–] 25 points 1 day ago (8 children)

It cannot be said enough: AI has never, ever made money.

It's excited people who don't know how it works, are gullible enough to believe it can replace all of their labor, and already had a lot of money.

The investor capital has been swirling around like the ouroboros for years now. Big numbers in the news ≠ profit.

  • source
  • hideshow 8 child comments
  • [–] 12 points 1 day ago (7 children)

    AI has been bobbing in and out of winters for 50 years. The big difference this time is that the winter will be taking the global economy with it.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 4 points 1 day ago (6 children)

    It hasn't though. The term has been used forever, but LLMs are less than 10 years old. This is new territory.

  • source
  • parent
  • hideshow 6 child comments
  • [–] 11 points 1 day ago (5 children)

    I genuinely don't understand what you're trying to say. This is a well-documented phenomenon over the last 50 years.

    Roger Schank and Marvin Minsky—two leading AI researchers who experienced the "winter" of the 1970s—warned the business community that enthusiasm for AI had spiraled out of control in the 1980s and that disappointment would certainly follow.

    As I said in the post that you categorically dismissed, the only thing that's different this time is the stakes will kill the economy rather than AI research.

  • source
  • parent
  • hideshow 5 child comments
  • [–] 7 points 1 day ago* (4 children)

    I understand that you probably think I'm only referencing the "hype cycle", and that's a fair interpretation.

    LLMs are significantly different from earlier versions of "AI" (non-consistent naming) throughout the period. These are not expert systems. These are a new beast which effectively did not exist before 2017. The only reason why LLMs are included in the group is arbitrary naming.

    I question the utility of folding them into the same hype cycle considering they have also already had more impact than any of the other cycles. This is new territory, and such arguments only take yet more attention away from the massive issue affecting the world at the moment. I would argue that you would be hard pressed to find another example of this amount of disruption.

    That said, I do not at all disagree with this:

    the stakes will kill the economy rather than AI research

    So I'm happy to just agree on that. Enemy of my enemy is my friend kind of thing.

    No hate at all. Hope you don't take it that way. I simply think that framing the discussion as "AI" (loosely defined as multiple technologies arbitrarily named AI) takes away from the clear and current issues around LLMs, especially if the argument is ultimately "just wait, we're in a cycle".

    This affects my life in a deep way right now. I couldn't care less about a cycle, in any event. (And that's on me, 100%.)

  • source
  • parent
  • hideshow 4 child comments
  • [–] 12 points 1 day ago (1 child)

    I think past AI winters is a somewhat useful analogy, but yeah, I agree, drawing too strong a comparison is misleading.

    Likewise for people comparing the AI boom to the dotcom boom and bust. Ed Zitron has ripped into this comparison repeatedly. Basically, the fiberoptics laid by the dotcom boom where useful for decades, whereas heavily used GPUs fail in ~6 years.

  • source
  • parent
  • hideshow 1 child comment