Like the cartel will allow the prices to fall even if there was more RAM available.

you are viewing a single comment's thread
view the rest of the comments
[–] 33 points 2 days ago (17 children)

That's a massively optimistic prediction if the AI economy doesn't implode. It's quite likely consumer electronics will be a faded memory (no pun intended) before they let people have their own devices again. I'd put it at 2035, with 2030 at the earliest.

If course if AI goes belly up, there could be a massive sell off and prices could drop as soon as next year... assuming there's a global economy left to buy electronics at that point.

How would you prefer your tech apocalypse? Matrix or Mad Max?

  • source
  • hideshow 17 child comments
  • [–] 11 points 2 days ago (6 children)

    Based on how markets responded during the AI stock selloff last week, an AI crash may not actually bring the the broader economy.

  • source
  • parent
  • hideshow 6 child comments
  • [–] 3 points 1 day ago* (last edited 1 day ago) (3 children)

    Judging by my experience of the previous 2 big Crashes (the 2000 one and the 2008, when I was actually in Finance), it's not the first big market correction that's the "oh shit" moment when basically everybody holding the overvalued stock just gives up and sells in a chain reaction were the more that gets sold the more the price falls the more people panic and sell.

    (I remember that shit happenned in 2008 such as the price of Bear Sterns falling massivelly and it being bought for pennies by another bank and that wasn't the "oh, shit" moment, that moment was a week later when Lehman Brothers went bankrupt).

    I suspect either we get a couple of cycles of big falls and small recoveries until prices fall enough so that most of the money that entered during the bubble is at risk of being underwater (and if you look at the stock prices of even the ones which fell the most, they're still way above the start of this year so most stockholders are still comfortably above water) - which if I remember it correctly was how the 2000 Crash started - or a big name bankrupcy like in 2008, and then the panic starts.

    When the real "fun" happens you'll see things like stockmarket automatic triggers for stopping the trading of certain stocks (which only happens for falls of the stock's price in a single day which are above a certain percentage) being frequently triggered all over the place and this stuff hasn't happened yet for big names.

    Something like Anthropic or OpenAI going bankrupt (possibly even already being bankrupt but hiding it with accounting tricks Enron-style) would almost certainly start the fireworks, though the way things are maybe Oracle is the one that blows up first and starts the run on the markets.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 3 points 1 day ago (2 children)

    My money is on Meta to be the Lehman Bros that kicks off the crash. I feel like they're over-extended and playing games with the math to keep from saying it out loud. Oracle is a possibility though, my gut is just saying Facebook is the sick man that's going to fall first.

    Disclaimer: This is not financial advice, nor do I claim to have insider knowledge of Meta's operations.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 3 points 1 day ago* (1 child)

    Personally, I'm pretty sure my schadefreude will be even more epic if Meta blows than if Oracle blows.

    But yeah, from the massive losses of the AI AND the Metaverse investments from Meta that I read about, it might very well be that they have unbalanced the company's finances too much to the point that a retraction of the money markets due to uncertainty around AI (making them unable to get loans or roll existing loans in good enough terms) ends up toppling the company.

    I mean, Lehman Brothers toppled essentially from over-exposure to Credit Derivates, a kind of financial instruments people outside the business knew had never heard about until it they started blowing up and all of a sudden nighttime public TV news were even talking about CDO^2^

  • source
  • parent
  • hideshow 1 child comment
  • [–] 3 points 1 day ago

    Exactly. If I understand the market signals, Meta has put a lot of money into datacenter shell companies that are facing a lot of local market resistance. If the market pulls back from throwing good money after bad on failed datacenter projects, combined with a lot of contraction in the social media space and slipping ad revenue, AND the mounting lawsuits, I feel like Meta could be the weak link. Again, don't quote me on that, I am guessing, but I feel like there's just enough publicly available info to make a circumstantial case.

  • source
  • parent
  • [–] 7 points 2 days ago* (3 children)

    If the bubble pops hard enough, the prices of those servers will drop so low they'll become consumer electronics. Oh that RAM won't fit in my desktop? Screw it, I'll figure out how to put Linux on that rackmount.
    (Wait, this this thing will only run FreeBSD? Eh close enough)

  • source
  • parent
  • hideshow 3 child comments
  • [–] 6 points 2 days ago (2 children)

    I'm not sure consumers are ready for a 1kWh server as a home/office item

  • source
  • parent
  • hideshow 2 child comments
  • [–] 7 points 2 days ago* (last edited 2 days ago) (1 child)

    All those data centers will need to be used for something. They'll either figure out how to use the AI optimized RAM for VDI, or the fabs will switch back and only sell to data centers.

    They'll sell us dumb terminals and a virtual desktop subscription. It would be great for like 90% of consumers if the companies running it weren't evil.

  • source
  • parent
  • hideshow 1 child comment