Post

can you give me some more specific?

📌

📌

I see Meta pivoted from burning insane amounts of cash on the Metaverse/VR into burning insane amounts of cash on AI successfully. At least with VR they were selling games to people even if they weren’t coming close to making money on it.

i read the whole thing. im a hater now

Eventually Jensen will need governments to buy GPUs, hence “sovereign AI”

Brilliant!

AI is great for helping the dumbest people in the office feel smart. Many employee who were once unable to contribute are now “contributing” heavily LLM-based work in the way that my 3 year old nephew “contributes” by pushing his Fischer-Price vacuum around the living room

kinda wanna go catch a Deepwater Horizon Devils home game now.

📌

Thank you!!

man i dunno. fucking idiots i guess

i think Samsung made it

If that functionality becomes mainstream, people will get their bank accounts emptied - because if your AI has access to an API that will give your credit card details to a restaurant without (non-AI) confirmation, any 3rd-party website can steal your credit card.

LOVE THIS

I'm hopeful that it becomes an employees market as jobs list to AI need to be filled. But agree that this turmoil was avoidable and stupid.

Edit: affects beyond*

This is a good visual. The servers that dot com companies required.

Milo Minderbinder as a service.

This is either extremely bullish in its resilience (massively profitable companies with massive cash piles) or extremely bearish in its concentration. Not entirely sure. Neither is the market.

Effing wild conflagration of money. Just spending beyond imagining.

I 'member when GPUs were for playing video games

it’s AIG all over again

The Gamers have truly won….

Ask ChatGPT: “Is this sustainable?”

Some argue that Amazon - and Amazon Web Services - were initially money-losers, and I have written a comprehensive explanation as to why it's totally different. AWS was quickly break-even, costing about 10% of Amazon's 2024 Capital Expenditures to build. www.wheresyoured.at/the-haters-gui/

Sort of makes sense why the stock market 📉 hasn’t been more violate given the Tariff King’s lunacy.

it's the ouroboros market

did Amazon ever have negative margins? I assumed it was always capex, etc, not margins that made the business red.

there has never been a point in my adult life when the phrase 'get while the gettin's good' hasn't been a guiding truism and nothing in my childhood prepared me to be that ruthless.

📌

Sorry, 19% of the 35% or 19% of the total? Either way. Wtf.

Started with crypto - use GPUs to make fake money for Ponzi schemes Now it's use GPUs to make a failing company. I can only assume there's a lot of embezzlement going on, because the alternative is too stupid

So to put this in terms a dummy like me can understand: “Big Capitalism just admitted that there is a Graphics card bubble like there was a housing bubble.” Correct?

Gosh, if only something useful could be found for that half a TRILLION dollars. We’ve failed as a society, miserably.

It’s an investment bubble like dot.com- there’s no incremental value behind it except in very specialised areas where relational database enquiry has been applied for years

In the data centre industry and can’t fathom how they are monetising AI the set up and ongoing operational costs are enormous. It has a price point above which most users won’t go. Not sustainable and the product…..not convinced, I mean are you really using it every day.

📌

I feel like I'm watching The Producers but instead of putting on a bad musical, they're boiling the planet & stealing our data 😭

„There is an overwhelming condescension that comes from fans of generative AI — the sense that they know something you don't, something they double down on. We are being forced to use it by bosses…“

this is a great piece - thanks.

Please check out stopgenai.com We are a grassroots, radical, political activism group.

ask any AI company salesperson about their competitors' products and they'll tell you a list of shit that's wrong with LLMs in general, including theirs. "It is difficult to get a man to understand something, when his salary depends on his not understanding it."

Feels to me like the novelty of AWS was dynamic scaling of allocated resources, which was as much a result of improvements in virtualization & containerization than any novel tech from within Amazon—the market was so established as to be practically commodity! It was just a hassle.

Meanwhile ... www.theguardian.com/technology/2...

But, Murati's Thinking Machines Lab is a company with no product on the horizon *at all* raising $2B with a $10B valuation. Of course, Nvidia was part of that funding round, so more in-breeding. Without IP infringement, it's tough to imagine any quick time-to-revenue/value. Mass insanity!

Just ask this genius.

Loved it

The implosion has already started, and the first big name will collapse spectacularly by the end of the year.

Still reading so not sure if you account for this, but all of these services require a lot of physical infrastructure. And big cloud computing providers are running out of "real estate" as we speak. They are actively turning down money from potential clients since they can't acquire enough

I wonder what they would have to charge in monthly fees to actually be profitable?

And that doesn’t take into consideration new Usage models vertically and horizontally in the services stack and industries. Same for competitors like salesforce who will see adoption in 1P/3P products in marketplace.

I’d love to see an investigative documentary about the similarities between Multilevel Marketing and the current state of “AI”.

At least tech wizz kid Peter Kyle MP (sadly my MP) signed a memorandum of understanding with OpenAI to increase the efficiency (take all the data) from the UK government yesterday 🥴

📌

Coward. (/lh, I'm sorry someone had to.)

14,000 words, every one of them worthwhile. (Also, I can’t wait to carry out that Open AI-recommended MLB road trip!)

@carlosenog.dev

📌

This is such a great fucking article and I need everyone in my life and their lives to read it, goddamn

"They say you gotta spend money to make money, well I dunno where we went wrong! We spent ALL of our money!" -Tom Haverford

I’m just sorry Douglas Adams isn’t around to see this…

I love that you write books and call it a newsletter

"they're in therapy." Looks like they took your advice! 😏

youtu.be/TZOoT8AbkNE?...

📌

Thank you for writing this. I haven’t finished it, but has been great so far. I am in tech and have felt much more skeptical about sustainability compared to others in my network.

I miss when gamers and PC nerds were the main audience for GPUs. The world is out of balance this way

GenAI is not made for money. It were brilliant experiments by one-person devs, like Deep Dream, GAN, First Order Motion, 3D Ken Burns etc. Sure, big tech comes like flies feeling money flows. But to see _just this_ is hypocrisy. Throw away your iPhone then, this sleak capitalism tool and icon.

I love being a hater I love being a hater

📌

Haters guide? Pump this into my veins

New update! In a call last night, "I am leaving you a legacy!" Oh my God how do I really tell him this is a fech bro ponzi scheme. He bought a GPU provider and earned 35%. What did the company that hardware went into earn? Oh it has been -17 for three years. This pyramid scheme won't last.

There is no ethical use of generative ai. The end. Don't even need to read the article but I probably will later anyway.

It took several years (decades?) for the computer industry to become profitable without relying on government contracts. Artificial Intelligence has already demonstrated its usefulness in certain contexts and costs will decrease over time. There may be a bubble, but generative AI is here to stay.

Every CEO and their mothers are in a rush to "slap AI to that", hence the bubble. Gen AI has its use cases, but you cannot expect to slap AI into something, and it just works. That's also why OpenAI is getting into consulting: "we will find the use cases for you".

I can hear you cackling: apple.news/AxaOODuxDTjy...

Well yes, there's all kinds of data. Capex before and after, Nvidia sales data, estimates based on the known requirements of other computational tasks .... the amount of capex devoted to LLMs isn't really disputed.

reading this makes me want to start an ai company, tbh.