- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
You must log in or register to comment.
🎉
Can’t happen soon enough. Fuck Sam Altman and his bullshit hype.
My previously popped popcorn may have gone stale, but I will surely have a new bucket ready for this.
We still have 2-3 years IMO. Investment momentum hasn’t dropped enough yet, and a lot of money going to data center construction is going to result in actual construction. There’s enough slight of hand possible to hold of the bust for a while. Well… There was before tariffs.
People have been predicting the edge of the bubble for almost a year at this point, since the week after GPT 3.5 came out. Been wrong 100% so far.
People are shitting bricks after that AI study from MIT saying 95% of AI projects fail.
I would wager 90% of AI initiatives are asinine wastes of resources pushed by people who do not understand them though.
Sure, but 95% of new businesses fail as well. Sunk costs are a thing, and people are planning around the savings of layoffs and letting LLMs do it all. AI projects can fail, that’s also a learning curve. In some cases companies will wait in hopes of AI rising to meet their needs in a year.
There is always a bubble in the tech world. People hype up a technology to drive profits to make up for the lull from the last bubble.
This doesn’t mean the death of AI, it’s just the inevitable settling of the market into where it should be. AI can do a lot of things decently and people are still trying to figure out what are all the things it can do really well. Some people made bets on things that it can’t do very well.
The next step is to expand on the things that it can do really well and plug it into user friendly experiences and UI’s. It’s not going away, but it will continue to evolve.
LLM’s are not the end all be all.
I don’t think the article is asserting that AI will die once the bubble pops. Just that the bubble is getting close to popping and there’s gonna be a shit show when it does.
Fair. There will be a crash and a leveling out with more clearly defined use cases and products. OpenAI already added 4o back as an available model. They’re gonna do what the market wants. It’s highly competitive.
A projection of AI’s development over the coming years published by something called the AI Futures Project under the title “AI 2027” states: “We predict that the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution.”
Every time I see a reference to this I’m reminded of the book: “The Third World War” by General Sir John Hackett. A speculative fiction that seemed to present it as if it were fact.
It didn’t happen this way, has yet to happen (obviously). AI 2027 is just a shiny modern version of this book crawling with some wildly speculative takes.
Everybody has been saying this. This wont happen, because of the “hype” death. This will happen when somebody in Open AI makes a major blunder. Shareholders and investors know “AI” is projected to have a very slow ROI, but that depends entirely on how useful it is. If it wont find use in any major HUGE branches (Surveillance, Medical, Investments, Science…) things it was marketed to do and help with THEN it will crash.
Part of me’s already got my SICKOS t-shirt on, going "Yes! Ha ha ha… Yes!!
But then, the other part’s like: “Frito! That’s your car…!”
For all of you just dying to see this bubble pop, do you realize how much money is tied up in AI speculation? That’s money that is, for the US at least, propping up our economy when it would otherwise be sliding.
Don’t eat the popcorn just yet. When this shit pops, you’re gonna need all the food you can get.
Burn it the fuck down. Magats need to feel the devastation they’re causing.
I’m 100% behind you, but the public will never relate AI and politics, not unless the datacenter is tearing up their forests and driving their bills. The relation is not simple conservative “not until it affects me” thinking. On this very particular subject, most Americans are totally ignorant. Do not mistake our attitudes and understanding on lemmy with what most people perceive.
Slash Medicare? Yeah, they’ll get that, eventually. One would hope. But AI? They’ll never connect the dots unless it hits their region, and even then, “Why my power bill so high?”
I’m ready for Great Depression 2.0. That event drove 60-years of Democratic policies. Yes, WWII was a major factor in our prosperity, but still, we had unions and social security and work programs and taxes on the rich and none of that was in question. We understood that was solid policy.
tl;dr: Americans are hopelessly ignorant on just about any subject you care to name. That’s not a put down, merely fact. Most of us can’t read 6th-grade level or higher. I was reading at college level in elementary school, though that must be bullshit. Nope. True.
Serious amount of upset AI boosters in the comments on that article.
Article entirety is chatGPT 5 sucks.
OpenAI has inside politics on being backed for Skynet. MSFT has put a lot of eggs in their basket. Sonet 4 is probably best US AI. Beyond OpenAIs incompetence and overhype, US AI is doomed by policy to ban solar/renewables. Every coal plant that was closed due to renewables growth was a massive quality of life improvement for every living being within 100 miles, never mind the climate destruction.
While datacenter growth in US was a bigger component of Q2 GDP growth than consumer spending (0.8% vs 0.7%), US does not have the supply chain to support electricity growth other than behind the meter solar. Refurbishing closed coal plants to power US AI is the low hanging fruit that doesn’t get to desired Skynet.
Deepseek v3.1 just dropped. Has big benchmark gains. OpenAI obviously cheated at benchmarks with v5, and its 120g open source model (training on them). I have no idea if deepseek did, but doubt it. Deepseek is also optimizing for training on Huawei chips (FP8 precision). Open source, locally run, is certainly only candidate for trust, regardless of China vs US source, and datacenter model somewhat doomed other than the to-be-well-financed Skynet ambitions.
Expensive monopolist extortion energy pricing in US, about to get scarcer, makes any industrialization or privatized Skynet militarization of AI an uncompetitive impossible goal. Extortion of consumers will cause resentment of GOP. Nevermind coming stagflation and collapse.
To solve those problems bigtech allied to fascism and they’re bringing freedom/democracy to latin america once again. 🇺🇸
But this is fascist oppression!!! – Angry Bolivians on Twitter
We will coup whoever we want to – Mecha Hitler’s boss.
