Edit, because I lol’d

Oh btw, I just remembered that I wanted to mention this here.
I actually didn’t find this thread through the forum website, but somebody linked it as a news source on the Fediverse/Mastodon.

We’re a reputable source now

Yes, for it requires intense physical training for a proper casting of “Fist”

Even those only whisper in hushed cants about the before times, when every inscription was carefully annotated in scrolls that could become as long as entire towers

FREEZE!

Also from HOMM3, it’s the tower hero sprite during combat

Apprentice mages usually begin by aiming at their own feet and yelling the incantation, some immediately turn to aim at the sun “just because”

Easy, it was another, greater mage. It’s mages all the way down

I’m surprised no governement, federal or state of any country, has started an investigation, since LG tvs in meeting rooms are common

The post using NES as a comparison is the main problem, in my opinion, especially given that Quake in 13k of javascript exists (only 3 levels, but still), which is a closer apples to apples comparison.

Besides, “needs to run on a multitude of browsers” is not hard unless you decide to make your javascript do absolutely everything, and even then it’s easy to keep it working on the major browsers, you just aim for the lowest common denominator (usually Safari). The browsers do the heavy lifting of running on a multitude of OSs and architectures.

I have an old (not quite 2013 though) Atom

That’s not mid spec, that’s bottom spec

So? That’s pretty much always been the case.

Ah yes, transistor fabrication has always been very close to the atomic level , no wonder it “pretty much always been the case”. /s

Consumer hardware is perfectly capable of running much larger models too if you give it enough RAM, and is only held back by RAM being expensive.

And average users having, at most, 8GB of RAM, with 2GB being eaten by Windows and another 2, minimum, by Chrome. Remember, average user is the person who doesn’t bother with latest games and doesn’t run linux, they only do some office stuff, watch videos and browse social media.

But prices for RAM will eventually come down again and hardware in general will keep becoming more capable.

At best, I expect RAM prices to drop back to their sane price somewhere in 2029-30.

Also, you may not have noticed, but for average users (people that mostly just browse social media and watch videos), mid-spec computers from 2013 are almost perfectly serviceable today - just swap the HDD for a SSD and you’re golden. Smartphones have also stopped getting significant hardware upgrades, despite shittier software trying to force obsolescence on older, mid-spec models.

Why would hardware stop improving exactly now?

Because current fabrication process of chips seems to be very close to limits imposed by the laws of physics. The 1nm fab process that TSMC might manage to pull by next year is possibly the literal limit of MOSFET. Unless someone comes up with a different fabrication process that can reach atomic precision, or somehow get more power efficiency from the same area/volume of transistors, or some wonder material, there’s no reason to expect hardware to keep improving at the same pace it did decades ago. “AI will discover-” It won’t. There is nothing indicating it might, much less will. I’d love to be proven wrong, but I seriously doubt AI will magic out some new way to get Moore’s law back on track

Accordingly, the AI models shipped with consumer devices will keep becoming more powerful and online models will be needed less and less for “average person” tasks.

I seriously doubt that for a single reason: money. Why let the user have the model when you can rent it off to them? Or, even worse, the model runs locally, but you have to pay a fucking subscription in order to use it because fuck you. “You will own nothing and be happy about it”

The reason why AI companies are spending so much money is to rollout data centres to run the AI in the first place.

Yet all companies fail to point whether there actually is, or will be, enough demand for ALL the capacity being built. Remember, we’re talking like 10+ huge datacenters that are effectively useless for anything other than AI.

If the AI industry has to pay off 5 trillion dollars over the course of 20 years, that’s 250 billion dollars a year

The author compares with Google/Alphabet revenue, a company which deals in ads and several other services. Guess the problem of comparing a company with a somewhat diverse portfolio vs. one that has one single product. Microsoft had a gross revenue of 281 billion dollars in 2025. AI companies need Microsoft levels of revenue by the end of next year.

Profits Can Grow, By A Lot

They can also not grow, because Anthropic and OpenAI can end up scaring customers into open models.

That means that there will never be a point where a company wins the AI wars

Wrong, Nvidia has clearly won the war by selling shovels that are useless for anything but AI. RAM manufacturers are also laughing all the way to the bank.

the average person isn’t going to use a locally hosted model, ever.

True. But average person can stop using said model if it starts being a scummy piece of shit charging hidden fees or eating up tokens more than the week prior. Besides, someone might try to rent their AI stacks for a profit and undercut the 2 inflated giants.

There’s Still Plenty Of Capital To Raise

Everything’s pointing to the opposite for 2027, with capital drying up, especially with “scaredy investors terrified of democrats being against datacenters”. Anthropic going public is likely due to its less than ideal credit rating (thus forcing it to deal with higher interest rates), so they can raise more money and leave suckers holding the worthless share bag.

Alphabet, Meta, and NVIDIA are all AI companies

Alphabet is a “everything about google” business that has a side gig with AI, Meta is a spy operation business that has a side gig with AI, and Nvidia is an overpriced hardware business. None of them is an “AI company”. The main problem of this bubble is absurdly excessive investment into AI, lots of scummy “it’s technically a loan, but legally the company can omit this from the balance sheet” shenanigans and the desperation to force AI down the throats of the whole world.

I don’t think the rhetoric about the AI bubble is really about finances. (…) I think the supposed AI bubble has become such a large talking point because it represents the hope that things can go back to how they used to be. (…) So, people wishcast. They tell themselves that this is all a giant bubble, and one day soon, everything will crash down and immediately go back to how it used to be.

Boy, imagine seeing an unproven technology receiving 2 FUCKING TRILLIONS OF DOLLARS, being promised another 3 in investments and thinking “nah, this ain’t about finances”.

As many people already said, the genie is out of the bottle. Nobody realistically expects shit to come back like it was before. Slop won, everyone that used to produce content lost, congrats, enjoy the shitshow of razing what little of the non corporate internet remained.

Even if AI turns out to be a bubble, it will be a bubble in the same way that the internet was a bubble, where the underlying technology remains useful.

The bubble is mainly in the datacenters, whose sole reason of existing is feeding AI. Remove AI companies and you’ll have literally zero customers. AWS, Meta, Google, Microsoft have no use for nvidia chips for actual server work.

Oh no! Anyway

“Yours truly, a 17 year old from India that thought they were doing something good.”

If this is true, I can cut some slack on the person being really, really fucking naive and acting on impulse instead of doing the bare minimum of talking to the actual authors.

“Why did I charge for AI Slop audiobooks? I doubt anyone knows how expensive it is to run a server like this, especially with the amount of usage this site got. It had 126,592 unique users within 24 hours of going up, and had terabytes of bandwidth in usage. In an attempt to keep the site up, I thought audiobooks were a good option. I sold like 11 audiobooks, if it’s any consolation.”

This shows how little this Ganesh thought things through. “It cost so much money!” - nobody affected cares, you acted the same way the abhorrent big tech companies do, slurping content en masse while giving the middle finger to the community responsible for creating and maintaining it.

Let’s come up with some fun fake statistics, all with pretty YoY graphics.

• Commodity prices of stuff that doesn’t exist
• Commodity prices of actual commodities (copper, steel, wheat)
• Stock prices of real and made up companies
• People with fursonas and their buying habits
• Rental prices of underseas landfills
• ROI for hotels investing in seagull tourism

TLDR - chatpgt got gamed by itself, as it saw the raw number of statistics and listicles listed on the page and absorbed absolutely everything.

I bought a brand-new domain, stateglobe.com, and generated 50,000 statistics pages in a single day with GPT-4.1. The site purports to provide stats on SEO, content marketing, social media, e-commerce, and web technology from 200 countries. Except every number was totally invented. I never checked a single one. I pushed them to the site with a basic AI web design template. This cost me around $30. Later, I added 15,000 listicles on top of the statistics. That was about $20 more, so all in all, this is an experiment that came in at the very low price of about $50.

[7 months later] In the last 30 days alone, the site got more than 600,000 requests from bots. Around 72,000 of them came from ChatGPT-User. That is the agent ChatGPT uses to open a page in real time while answering someone. So these are not training crawls. Each one is ChatGPT fetching a fake statistics page to answer a real question.

On October 7, stateglobe.com was ranked second for visibility on these prompts, behind only data powerhouse Statista.

I wonder if that’s also applying to countries that still have heavy usage of FB, mostly SEA ones, from what I remember.

Really good read. Making the car comparison and all its effects hammers (pun intended) down that “just a tool” can be an incredibly complex thing that needs a lot of variables to work properly. A hammer is a simple tool, a computer is not simple - there’s a global logistics chain that allows their existence - and neither is AI.

We need to fight back against design that flattens our humanity at the benefit of efficiency and productivity. We need to question how our tools have created infrastructures and landscapes that are hostile to human existence.

Emphasis mine. I’ve been thinking about that a lot. Everything is about being more efficient, more productive. Fuck that shit, I wanna enjoy life.

I guess we’re too used to melon musk and trump meltdowns, so anything that isn’t balls to the wall incoherent rambling is deemed like a normal conversation on xitter

Aw come on, being UTTERLY ANNIHILATED isn’t that weak :(