The author could have made that argument, if she had wanted to, but the actual piece as written does not make it. There’s very little discussion of AI being overhyped in some concrete way, and quite a lot of discussion of various bad deeds committed by “tech” and “tech billionaires”.

I don’t think the “GPU bubble” will burst, because linear algebra is widely applicable. It’s no particular mystery that GPUs have found universal applications. I feel sympathetic towards your first paragraph, but there is no conspiracy behind the success of GPUs.

Anti-tech critics love to cherry-pick overhyped tech from the past and then pretend like it was the only technology investments occurring. It's always the same story about crypto, NFTs, multiverse, then they draw a line to the new thing.

Rewind the clock further and the contrarian play was to talk about how WebVan and Pets.com failed, proving that internet commerce was a fad that was going away. There were so many identical stories about how dumb investors were to be spending money on e-commerce after Pets.com and WebVan proved that nobody wanted to shop online and that delivery was unworkable.

More recently I remember the endless stories about how ride sharing was going to fail and Uber and Lyft were going to disappear after the VC money ran out. There were blogs just like this one predicting that those dumb investors were going to lose all their money on such a stupid idea.

This type of contrarian reporting always operates on a sliding window of recent failures, trying to convince you that the current thing they're on about is identical to past failures

These articles get traction on HN, but when I read them there isn't a coherent argument inside. It's just a collection of different headlines and stories meant to imply that AI is bad across the board and nobody wants it, but there isn't an argument being formed. It's appealing to those who already have the conclusion in mind, but there is no convincing argument in this post

That’s fine but it’s about perception, argument and sentiment. It’s not about the actual technology, which has the final say.

i can’t believe i’m seeing an “yet you participate in society. curious!” reply in the wild

we both know there is an obvious difference between “using Ghost to share a hypertext article” tech and “the metaverse and palantir” Tech

Seriously? Having a blog is now same as dissolving your brain in anxiety driven doom scrolling?

Big tech Arpanet and CERN are full of it!

Big Tech is not the same thing as all technology.

I agree with you on most parts, but first, not everyone can actually run a model on your own computer that will do 98% of what you need - for most people it's rather around 40% of what you need, depending on the exact model and method of usage; and second, there's still these 2% remaining. And don't forget that not everyone (I mean, not every casual user) can do local inference just due to the lack of technical skills.

And by the way, who will produce this "model that you can run on your own computer"? Although, to be honest, right now in most cases that's not Meta or OpenAI anyway, even though they have some (outdated) open-weight models.

> I dont understand how they will all be profitable when you can run a model on your own computer for free that will do 98% of what you need.

For the same reason how movie companies are reaping record profits from streaming services - convenience and fearmongering.

"You wouldn't download an AI, would you?!" or "Think of children and elderly!" or satanism, or billions of other tricks they already employ to keep you docile.

Also, majority of people do not have hardware to run models that compete with LLM vendors, at this point I'm pretty sure majority of people don't even own a PC anymore.