“reading” one of those pages but don’t remember what the page was about.
That’s one of the biggest tells of AI-written text. It uses a lot of words to say very little, but does so in a very authoritative-sounding (or needlessly flowery) way.
SEO-based business models used blogspam before. It’s the same SEO garbage that gets it into search results, but the content is now AI slop instead of contracted labour at pennies/word.
And search is garbage, now, because of enshittification; Google gets more money when you give up and couch the sponsored links, and re-query or load more pages of results to load more ads. So there’s no incentive for them to filter the spam.
This article isn’t about AI companies at all. Not one is mentioned, iirc.
This article is about “regular” businesses’ use of AI.
OpenAI has a snowball’s chance in hell of surviving. LLMs cost a lot of compute. They’re burning through cash. Operating costs are more than double revenue. Their net operational losses are about $1 million USD every 40 minutes.
And somehow they’re trying to put half a trillion USD into building more datacentres to make even more advanced models, which will be even more compute intensive.
Meanwhile, as venture capital has been committed to a whole series of AI companies and data centres, venture capital is dying up. Nobody has gotten a payout yet, since there’s no path to profitability for any of these companies.
And it’s coming to a head this fall, when OpenAI needs to pay their suppliers for the expansion they’re building, and there’s no reason to believe they’ll be able to raise enough more investment to cover their costs.
It doesn’t even take OpenAI failing, either. There’s so much debt (“leverage”) and circular cashflow going on in this space, between the AI companies, data centres, computer hardware manufacturers, and construction companies, that any one of them failing could cause cascading failures, like dominoes. Worse than the '08 financial crash, most likely.
So no. It’s not going to be like YouTube. YouTube is cheap to run, compared to LLMs.
And the worst part of it all: LLMs aren’t even very good! It creates an illusion of productivity, but it’s all bullshit, either doing a shitty job, or taking more time to prompt fondle than it would have taken to do the job by hand, or building up tech debt that’s going to make massive projects unmaintainable.
It has some use cases, sure. I use it almost daily, tbh. But only because someone else is footing the bill. It doesn’t produce nearly enough value to justify its costs.
A company borrowing money to buy it’s own stock makes sense though, doesn’t it? A share buyback reduces the number of shares, so remaining shareholders are holding more equity (as a percentage) than prior to the buyback. It’s just the reverse of issuing new shares. If the company has no productive use of the capital, then a share buyback is a way to make the company more “lean”, shedding unneeded cash to increase (relative) value.
Borrowing money to do so just means that they are deemed credit worthy by enough bond investors that they can borrow at low enough rates that the debt repayment costs are less than the value shareholders would expect from a dividend payment and/or that they don’t want to issue dividend payments for some other reason (like the idea that dividends should be consistent and only ever increase or they’re valuation gets slaughtered.)
The whole stock market is a bubble now, anyway, so this is the heart of our problems. About 2 decades ago, financial reporting allowed companies to shift their PE ratio away from Price-to-Earnings ratio and instead report on Price-to-Forward-Earnings ratio. This is the company’s projection of their future earnings potential, but investors just seen to accept that a “PE” ratio means the same thing it did for the preceding, like, century. A PtFE ratio of 25 is insane, on historical contexts, yet that’s completely normal now.
Insanity. And yet the market keeps going up. Even the '08 crash was just a small blip, compared to what it should have been.
I’ll just put my tinfoil hat back on over here.
I’m liking Krita and Photopea (web app), but I’m not heavy into photography. I haven’t looked for a Lightroom replacement.
It will also likely work really well, apparently. I think you just need to be careful to pick a distro that comes with NVidea drivers, like CachyOS, and it will likely just work. Test with a live USB boot.
If you have an AMD GPU and don’t care about playing games that require kernel-level access for anticheat (ew), then Linux might just work better for you than Windows, for most games.
Like, getting Minecraft installed and working with mods in CachyOS just required installing Prism Launcher from the CachyOS repos (1 easy step) then launching it. I didn’t even need to open a web browser to download an installer.
Heroic Launcher is amaze balls, too. It pulls all the free games I get on GOG, Epic, and Amazon (iirc?) into one library that looks and works like Steam’s (amazing) library. So slick. (I think it’s preinstalled in CachyOS, too.)
Not mentioned, but if there are mobo monitor connections, try those, too.
But yes, this is almost definitely a hardware problem since it’s also happening in Windows. The only other plausible option would be the hardware’s firmware, but that seems unlikely…
It could theoretically be an incredible fluke to have a software issue in both Windows and Linux… Maybe the same weird edge-case hardware interaction that’s the same between two versions of a closed-source NVidia driver? I can’t see that as plausible, though.
If OP is in a developed country, used monitors are cheap. My vertically-oriented side monitor I got for $20, and I only even paid that much because I needed one that could go vertical orientation without a monitor arm.
So do they. But then the tiny tool they built for fun keeps expanding as they add features until it’s useful, then really useful. And some eventually become a small, ignored, absolutely critical components in software used by millions. Too small or unsexy to get any money, but user errors or scammers or AI slop or bugs or feature requests lead to enormous volumes of email, comments, forum posts, vitriol, pressure, stress, angst, burnout, depression.
The challenge, as always, is to never underestimate a bubble’s capacity to outlast your solvency. I personally know people who have been heading against the housing “bubble” in Canada bursting since 1999. They’ve spent a lot of money with nothing to show for it, yet, and missed out on housing prices, like, quadrupling? Quintupling?
So, good luck. Buying out-of-market puts might be a safer bet, since you’re most likely to “just” lose all your money, with a small chance of a massive payout of it “properly” crashes.
Exactly right. Parent poster is conflating the investment in “AI” since 2022 (almost exclusively meant to mean LLMs, like ChatGPT) and specialized “AI” systems (almost exclusively “machine learning” systems).
A LLM is just about useless for any sort of surveillance or data analysis tasks. The bigger fear with LLM proliferation is as a propaganda machine, astrotufing the whole Internet with mass LLM-generated bullshit.
Yeah, I installed Enterprise edition on my desktop, which allows you to cut out all the bloat and spyware. But it takes a long time to do, and I’m not sure I got everything since Windows Updates can change anything.
That love of tinkering is why I’ve landed on not using an immutable distro for my first time installing Linux since the 00s. CachyOS is what I landed on; now I just need to catch up on work so I can take a day to tinker with my setup.
For context, I semi-broke my current Windows 11 install by trying to manually edit the registry to remove all traces of a piece of invasive, uninstallable bloatware (that comes direct from ASRock… the bastards) I accidentally installed. Turns out my sound drivers are from the same company, so when I deleted all entries with that company in the search terms, I FUBARed my Bluetooth audio and 3.5mm microphone. And didn’t backup the registry.
I like to tinker, and if I need to reinstall my OS anyway, so now is the time to finally switch!
Sure, but even so: that hardware isn’t going to be obsolete or wear out for a long time.
Compared with going to a concert that can be $10K for 2 people to attend a single show.
I could see it for some things. It would be great for browsing 3D print models or furniture—items where the shape/size matter a lot.
Like, an IKEA app that uses AR passthrough to show furniture options in your space? That’d be really cool, if it could be trusted not to spy on you.
Or an interior design app, that gave you a variety of options it pulls from a variety of sources, and you could add items to your cart right from the app?
For 3D prints, it would be great to have a virtual storefront of models to see, pick up, rotate, etc. 3D print .STL files are shockingly large (often 100MB+ for a single model), so idk if it’s realistic, but it would be a great use case if it was feasible. (My current special interest is showing…)