• Bloobish [comrade/them]@hexbear.net
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 days ago

    Love how the headline is misleading because it’s likely saying a recession will happen before 2028 so likely this coming year (or starting right now TBH)

  • CyberMonkey403@lemmy.ml
    link
    fedilink
    arrow-up
    0
    ·
    2 days ago

    Apparently banks are now claiming to have “technical issues”. Guess what they claimed before the 2008 crash, while they scrambled to pull out the money.

  • halfpipe@sopuli.xyz
    link
    fedilink
    arrow-up
    0
    ·
    2 days ago

    Of course it’s a bubble. Of course it’ll burst. Whether that’s timed perfectly to one party’s advantage in the 2028 election is anyone’s guess.

    Thing is, plenty of companies that we now depend on built their businesses doing exactly this: taking peoples’ money for a product that didn’t exist at the time, but would in the future.

    Netflix. Uber. Amazon. DoorDash. Postmates. Legal Zoom.

    There’s more, but the point is that every one of these brands used investor money to subsidize initial costs to customers who developed reliance on the product. As prices increased, customers’ dependence on steamable content, impulsive rideshare, two-day delivery, grocery shopping…you name the service, and people who incorporated it into their lives found a way to pay for it.

    But unlike these prior successes, with AI, when that price suddenly spikes, customers aren’t going to be able to afford it.

    • ☆ Yσɠƚԋσʂ ☆@lemmy.mlOPM
      link
      fedilink
      arrow-up
      0
      ·
      1 day ago

      There’s also another aspect to this which is that openmodels are becoming more capable and efficient all the time. They’re already good enough to do what most people need day to day, and if they get a bit more optimized where they can run on more modest hardware we’ll start seeing products focusing on running stuff locally on the user machine. And the whole rental business model will collapse.

      And building products around local models has a ton of advantages. You can advertise the product as being privacy focused, and you don’t have to operate server farms to support it. The one huge difference between stuff like Netflix, Uber, DoorDash, etc., is that operating these services is relatively cheap. So, the cost is largely negligible relative to the revenue. But running LLMs is actually costly, so the profit margins are thin to non-existent here.