Security Verification
www.ft.com
  • BananaTrifleViolin@piefed.worldEnglish
    6·
    3 days ago

    So there is a key message in this article which heralds big problems for the AI bubble: AI is being commoditised. I.e. AI users are no longer just going for the biggest/best model overall but instead looking for the “good enough” models for specific uses.

    This matters because the entire AI speculative boom has been predicated on the idea that the AI industry is not a commodity but a monopoly in waiting like Search or Social Media - one winner will take all, and so investors are betting heavily on their picks to be The One. But instead it looks like AI is more like a commodity - it will instead be a market of many players, with specalised models and differing levels of sophistication for different tasks. In that world, there isn’t going to be a “one winner takes all”. Instead there will be choice, and people will go for the cheapest model that is “good enough” for the task at hand. Do you need the most cutting edge expensive model to proof read an email?

    The crazy valuations of big AI companies was already a problem even if the idea one winner takes all was even correct. It would have meant investors were speculating on huge rewards for backing the right horse, and inevitably many investors were going to lose out as only one horse can be the winner. Everyone of course believes that their company is The One, and everyone else will be the ones that take the hit - typical of a bubble like this. But if AI is a commodity, then no investors are going to strike it super rich as no one company will be dominant and essentially print money for them in the future. So what you have a load of massively over-valued companies that will instead be bit-players in a big market, and never reach the income levels that justify trillion dollar stock floats.

    It’s not dissimilar to the dot-com bubble. The Internet didn’t go anywhere but the first speculative investors were burnt hard when the bubble burst. Other companies came along and built the world the first investors hoped to own. It’s looking like AI is going to same way - there won’t be one Google dominating everything, instead there will be lots of failures like boohoo.com or pets.com or go.com.

    I think we’ll see the big companies like OpenAI and Anthropic rush quickly to get their floats done, because sensible investors know now is the time to cash out. Float the company, quietly sell your shares and let others take the hit as it all comes crashing down.

    • scytale@piefed.zipEnglish
      1·
      3 days ago

      Instead there will be choice, and people will go for the cheapest model that is “good enough” for the task at hand. Do you need the most cutting edge expensive model to proof read an email?

      That’s already happening now, with companies telling their employees to be mindful of their usage and use lower reasoning levels for mundane/basic stuff (documents, summaries, etc.) and reserve the more costly models and reasoning for the extensive tasks (i.e. coding). The AI companies are now starting to charge money and suddenly tokenmaxxing isn’t such a good idea for management.