Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
Google AI Facebook Microsoft

Tech Giants Are Set To Spend $200 Billion This Year Chasing AI (bnnbloomberg.ca) 52

Three months ago, Wall Street punished the world's largest technology firms for spending enormous amounts to develop artificial intelligence, only to deliver results that failed to justify the costs. Silicon Valley's response this quarter? Plans to invest even more. Bloomberg: The capital expenditures of the four largest internet and software companies -- Amazon, Microsoft, Meta and Alphabet -- are set to total well over $200 billion this year, a record sum for the profligate collective.

Executives from each company warned investors this week that their splurge will continue next year, or even ramp up. The spree underscores the extreme costs and resources consumed from the worldwide boom in AI ignited by the arrival of ChatGPT. Tech giants are racing to secure the scarce high-end chips and build the sprawling data centers the technology demands. To do so, the companies have cut deals with energy providers to power these facilities, even reviving a notorious nuclear plant.

This discussion has been archived. No new comments can be posted.

Tech Giants Are Set To Spend $200 Billion This Year Chasing AI

Comments Filter:
  • by ebunga ( 95613 ) on Friday November 01, 2024 @11:42AM (#64912387)

    Headline should read "waste" not "spend"

    • I wonder if these "investors" really understand what they're buying into. Are the people signing the contracts & cheques doing sufficient due diligence & making sure they understand what GPT LLMs & the like can do & how they work, & most importantly what the already known limitations are, or are they just listening to hype & being wooed by the likes of Sam Altman? I suspect the latter & that this isn't going to end well for anyone... except for those who understand the opportunit
      • by Baron_Yam ( 643147 ) on Friday November 01, 2024 @12:22PM (#64912503)

        Our economies' histories are full of good things being replaced by crappy but much less expensive things. AI is going to replace good people with crappy code, but the price point on the final product will be so attractive we will choose AI even as we complain endlessly about the quality issues with it.

        • Nope. Not even that. So far, only highly knowledgeable & experienced coders can get LLMs to work more efficiently for them. For everyone else, it makes coding take longer, more buggy, & ultimately more expensive.
          • That's one use of AI. There are plenty of others.

            And I disagree even on that one; my whole point was that the quality drop won't be an important factor in the switch.

      • by jhoegl ( 638955 )
        So, the whole thing is a circular jerk of sorts in that they spend investor money on hardware that they can use in other instances besides "AI", and when the eventual colapse happens because the hype train will let everyone know that its really just a dumber search engine, then the stocks will fall, but they will have sold off stocks, and bought on the dip while also owning hardware that can improve the business in other ways.

        Its a stimulation of the tech industry that saw a huge fall from the 2020/2021 t
        • spend investor money on hardware that they can use in other instances besides "AI", and when the eventual collapse

          Similarly the dot-com bubble of the early 2000's paid for lots of internet cabling to the masses that was eventually put to use. It's just that the original investors didn't get a cut because the original companies bellied up.

          Railroad build-out in the late 20th century also built lots of tracks that were later put to good use, but the original investors got screwed by the bubble poppage.

      • They are signing into a massive automation boom that is going to replace tens of millions of jobs with computer software. Bare minimum we are looking at 7 million taxi drivers going away and losing their jobs and their livelihoods in about 5 to 10 years. The only thing holding it back is getting people used to the idea of a self-driving car and working out a few legal problems.

        That is worth trillions. The people investing in it are the ultra wealthy and they've set the system up so that they can't lose.
        • by jhoegl ( 638955 ) on Friday November 01, 2024 @12:29PM (#64912533)
          It absolutely will not do this in the next 10-15 years at least.
          • You've got Google and YouTube you can look up the videos of the waymo taxis driving all over Phoenix and San Francisco without drivers and having no problems doing it whatsoever.

            They don't do as well in wet weather but outside of Seattle that's not really a problem. And you can use a handful of gig economy workers to make up for when the weather isn't good enough for them to operate. Best case scenario you put 5 million people permanently out of work. And it's the bottom rung of society. They have nowh
            • by jhoegl ( 638955 )
              First, Waymo can be stalled by something as simple as a traffic cone. They are also easily stalled by people crossing the street. If you have a constant stream of people crossing the road , it will never move. I saw this at a stop sign intersection, the Waymo car was at the stop sign, people were constantly streaming out of a theatre and not pausing to allow traffic. The Waymo literally held up traffic.

              Second, these cars operate SPECIFICALLY on well maintained roads, meaning any road that does not have
        • GPT LLMs & other computational statistics models have nothing to do with actually controlling physical machines in the real world. In the USA, they're prematurely rolling out autonomous taxis & hoping that they can get away with it. They cite general traffic accident statistics rather than anything more representative, like professional driver accident statistics, like those from taxi & limousine drivers. It's just more sales-hype bullshit. Human taxi drivers aren't going anywhere for the forese
          • Waymo is going to destroy taxi cab jobs. That's going to happen and soon. When I brought up LLMs that was a completely different type of work. I'm talking about customer service jobs. LLMs are going to destroy tens of millions of customer service jobs in the near future. That's probably a bit further off than the taxi cab jobs but it's still coming.

            Think about how we fake the unemployment numbers by not counting people who stopped looking. We've got about 7 million men who have just given up trying to f
            • I think you're hallucinating.
            • On the other hand, AI will help people who are not that smart work in jobs that would normally require a lot more intelligence or training. So, maybe some retail workers can go into marketing, for example, because they can leverage AI to learn how to do marketing.
        • by ebunga ( 95613 )

          It's going to replace management consultants and marketing departments who are bullshit-driven anyways.

          Now that I think about it, AI isn't that bad.

          • You don't realize what management is for or that it was already eliminated some time ago. Management in America and other modern economies was created as a means of keeping an eye on unionizing employees and making sure there was somebody close to the employees representing the interests of the owner class.

            When we broke the unions under Ronald Reagan and Margaret Thatcher back in the '80s we no longer need it managerial classes. That's why they were massive purges and managers back then and if you actua
      • by taustin ( 171655 )

        I wonder if these "investors" really understand what they're buying into.

        They rarely do, especially with tech companies.

    • Okay and I hope you get your Funny mods, but the joke I was looking involves how we are getting chased by AI-driven spam. The clickbaity title of the story is asking for such... Especially getting chased political garbage spam just now. Should I predict a xitter-storm?

    • Ssshhh, my mortgage is funded by hype. It's the dirty secret of IT. Our "Sacklers" will go longer without getting caught because the mayhem is spread thinner and wider than opiate addiction.

    • 200 Billion is going to seem cheap compared to the cost when AIs start chasing companies - with lasers on.

  • by MpVpRb ( 1423381 ) on Friday November 01, 2024 @11:52AM (#64912413)

    The good news is that some of that investment will lead to really useful stuff
    The bad news is that most of it will be used to create crap generators

  • Next dot bomb (Score:4, Interesting)

    by sheph ( 955019 ) on Friday November 01, 2024 @12:10PM (#64912465)
    This is going to be the new dot com bubble. Everybody and their dog with a hair brained scheme throwing around buzzwords to secure venture capital to fund their new startup. When the music stops the results will be the same and there will be a small number of really good ideas to come out of it but most people will lose a bunch of money.
    • by Z80a ( 971949 )

      I find funny the many that are basically just chatGPT with a "please pretend to be the product we're too arsed to actually make" pre prompts

    • This is going to be the new dot com bubble.

      Let's not forget though that the dot.com bubble was built on a real and fundamental shift in how technology, business and society worked. After the dot.com bubble exploded, the internet had radically changed - from something of interest mainly to nerds and academics to something that touched everybody's lives and changed the way we live, communicate, entertain ourselves, shop and so on.

      I believe we're at a similar stage with AI: the current froth of startups, competing models, speculative investments and so

  • In a gold rush... (Score:5, Insightful)

    by Drethon ( 1445051 ) on Friday November 01, 2024 @12:16PM (#64912485)

    Don't hunt the gold, sell the tools.

  • ..."That means there has to be a pony under here somewhere! Keep digging!"
  • by xack ( 5304745 ) on Friday November 01, 2024 @12:17PM (#64912489)
    Again showing poverty is deliberate, as we have the technology, just not the politcal will to solve it.
    • by hwstar ( 35834 )

      It isn't political will.

      It's lobbying from the Business Lobby.

      They need a reserve army of hungry workers who are ready to take any job to house, clothe and feed themselves.

      Yes, I know this is from Marxism but it really does apply.

      https://en.wikipedia.org/wiki/Reserve_army_of_labour

      • That's how fascism works. The corporations take control of the government through bribery and then they CONTROL the political will.

        • Funny how when you go in either direction far enough you run into the same authoritarian bullshit.

          • I don't think "either" sufficiently addresses the number of different directions.

            • I guess it depends on how you look at it. If you check this out:

              https://www.politicalcompass.o... [politicalcompass.org]

              Pretty much any government is the authority and places them above the horizontal axis. If you go left or right from center you don't necessarily become authoritarian but on a practical sense, you end up there. Of course, going down on the vertical axis leads to freedom until it leads to anarchy, which is chaos.

              Some countries do manage to straddle the horizontal line but most tend to be one flavor or anot

    • No one gives a shit about poor people. If we did, we wouldnâ(TM)t be wasting time posting drivel on slashdot.
  • How many trillions will be wasted on this nonsense before everyone wises up and realizes they've been taken for fools?
  • by larryjoe ( 135075 ) on Friday November 01, 2024 @02:23PM (#64912837)

    "Three months ago, Wall Street punished the world's largest technology firms for spending enormous amounts to develop artificial intelligence, only to deliver results that failed to justify the costs."

    What does Wall Street know? Reading articles pretending to understand stock movement makes my head spin. Sometimes an article will state a reason for the market's rise, followed one hour later by the exact opposite reason for the sudden fall in the market. The articles are mainly to generate online clicks.

    The big question is when we should expect AI spending to manifest in corporate quarterly reports. Even if AI really worked and companies knew exactly what they were doing, there is a lag between greenlighting and impacts on quarterly reports. People have to be hired, equipment has to be installed, data has to be gathered, models have to be trained, customers have to be found, and money has to be paid. Is a reasonable lag 3 months, 1 year, or multiple years?

    Add to all this the fact that we're still in the research phase, so little of the above is happening right now. The current AI craze is based on potential and FOMO. So, the actual lag to quarterly report impact is likely years. For anyone experienced with this process, this timeframe is common for most new products.

    • Companies are using LLMs to generate corporate quarterly reports. Maybe they're tired of doing "work" and would rather just do "talk" so they're generating powerpoint slides. They're also asking LLMs to suggest where they should focus in the next cycle to improve the numbers. It makes sense, it's tedious stuff and I wouldn't want to do it all day. What insights do you think you're going to gleam from the OLAP Cubes this quarter? The data science people were already doing this stuff with ML, it's just a lo

      • Companies are using LLMs to generate corporate quarterly reports.

        Are there really companies that use generative AI to create corporate reports that are filed with the SEC? Since the CEO/CFO are legally responsible for signing off on the reports, are there any such executives who are willing to allow AI to create these reports that they blindly sign?

    • by khchung ( 462899 )

      What does Wall Street know? Reading articles pretending to understand stock movement makes my head spin.

      Wall Street knows how to manipulate stock prices, making it swing up and down to make profit. Your head spinning from reading their articles showed they worked. If you "invested" based on these junk articles, you would contribute to Wall Street's profits.

      For layman to make money from the stock market, look for healthy companies that pay dividends, buy and hold their stocks, profit from the dividends and never sell them without good reason. For noobs, buy index funds that pay dividends so you don't even n

    • by evanh ( 627108 )

      It's two years of hype now.

  • You can fix the USA economy with a 10 cent per kilowatt hour tax on power used to train AI. Maybe with 20 cents, you can also reduce the national debt.

    Seriously. There needs to be something good coming out of all this wasted power.

  • If these companies can throw this kind of money at AI make them build nuclear plants to power all the datacenters they’ll need. We can push AI back ten years to get a few plants approved and built.

    • by ebunga ( 95613 )

      They demand clean energy... from the states they "bless" with their datacenters as part of their tax deal while shopping around. They get their greenwashing and we have to pay for it.

  • Just on marketing, not actual AI yes?

"If it's not loud, it doesn't work!" -- Blank Reg, from "Max Headroom"

Working...