Post

Yup. It’s so not ready for real life complexity that isn’t intimately tied to something already simplistic or deeply code based. Ten years from now 👍, but it currently has zero use for me as I already have a fulltime job and do not have the bandwidth to be some outside actor’s unpaid test monkey.

That is a bad assumption

Who gives that man a mic. Give that man a job that contradicts his shittiness *only

This is going to be a worse capital misallocation than "all we can do is break down these ships for scrap"

But they also won’t be because there will be no reason to make new generations of ai compute

Bingo. It is much, much, much worse than the dot.com bubble. Huge circular financing of Capex that is being recognised right now in "earnings". So when the bubble pops earnings won't just stop growing, they will massively contract. It's the ultimate example of the Kalecki profits equation in action.

i love how “it’s just like the dotcom bubble” is seen as a defense. Just because we got cheap fiber or whatever doesn’t mean it was okay to waste all those resources in the bubble.

Love that the entire economy is turning into a furious self-suck session. Very normal.

Typically a building needs 10-20% annually spent in capital to keep in relevant beginning at about year 10 through eternity if the building is maintained properly. Not including opex to run the building.

Maybe they will be profitable in year 2030! (2030*2029*2028*...*1)

Fair points but specific to GPUs; they’ll still have a lot of utility come 2030. They’ve been a niche element (ex., VDI) in data centers for many years now. So beyond LLMs there are/will be use cases *Note, I’m not discussing the sheer volume of the build outs making sense, just GPUs as a processor

I dunno, if I had a business model which was "version 1 loses money, but version 2 will be great and make lots of money" I'd want to spend as little as possible on gear that will only be used on version 1...

No it’s not

Nah, they just let a complete moron write an opinion piece that will prove fantastically incorrect.

video games will run at 900,000 frames/sec tho

better climate models. new materials science. modeling protein structures. finding new medicines. image processing. like diagnostic images, or telescope images of deep space. finding patterns in large data sets, could be in medical records, or discrimination in criminal sentencing. lots of stuff.

I dunno. With all those GPUs you could play some really kewl video games.

Weird to bring up railways and fiber-optic cables when those things were famously overbuilt at various points and when the expected demand for them didn't appear the companies that laid them down collapsed. And in turn did very real damage to their countries economies.

I don’t believe he pays Forbes to host his blog

Step 1: it's not a bubble Step 2: it's a bubble but bubbles build infrastructure we'll use in the future (being said now) Step 3: it was a bubble. The money was wasted. But it was a reset and our economy needs these (can't wait for this one...)

I mean the railroad bubble in the US left behind a bunch of abandoned railways that no one used. States are still removing them when they get in the way. There are also plenty of abandoned factories.

of that fraction using generative AI- a portion is already dropping/abandoning it bcuz it sux, bcuz it drains job satisfaction, etc

"Today’s risks are different and arguably more recoverable" is the scariest thing I've read all week, yeesh

Great example of how there's truly never any repercussions for professional bad take havers I'm sure that this will continue when he's inevitably proven completely wrong about this take as well

Ed, you rock. I loved how shocked Ari M. was by your well-informed fervency. Keep being awesome, please.

The implicit assumption that there's no legitimate market need for drugs 🤯

Even fully-functioning data centers are useless shit if they’re surplus to capacity, since the hardware becomes rapidly obsolete and they’re expensive to maintain. It’s not like railroad tracks, which require maintenance but are basically 1850s technology, or roads, which the Romans perfected.

Other than wishing very hard, how does he know the 17.8% is the floor not the ceiling?

This bubble is much closer to the bitcoin/crypto compute bubbles, but 100x larger. Not excited to see these pop even though they will.

Until recently, The Guardian was one of my last refuge when it came to general press, but under the cover of pluralism, they become more and more an echo chamber for nauseating discourse...

This. Like when my phone power button was silently changed to summon Gemini and it got used precisely once to prompt "how do I change the power button on a galaxy s21+ back to a power button" Still makes me part of the 17.8%?

I'm shocked by him thinking 18% of people using it is low. What percentage of workers does he think needs an LLM? Personally, I'm thinking 18% is way too high, most of them just have it forced into their MS office and would remove it if they could.

I read so many articles in The Guardian, but skip everything by Gene Marks. His takes are just bad most of the time.

Data centers 'leave behind' outdated, inefficient hardware and acres of concrete that will cost more to get rid of than the land could possibly generate. (Side q: why use 75-year material to house ~3 year tech?)

Only 0.1% of people drive super sports cars, maybe they should invest there. Seems like we are only at the beginning of super sports car adoption.