lowsong
The entire AI market is propped up on the hope that the models will magically solve all of these problems, and now that even Altman is realising they won't he needs the slowdown to stave of the crash for as long as he can.
And yet we had the dotcom bubble.
> It doesn't matter that revenues exist (mostly fake and circular).
Oh people are paying for AI now, but will they do so when costs go up? Will they do so when the hype dies down and they realise that it doesn't provide the benefits it claims? Remember it's a feedback loop, a depression in the stock market will make companies want to cut costs to improve their numbers, and questionable AI spend will be the first thing to go. Once this starts to burst, it goes down fast.
> It doesn't matter that real unit costs go down (models are getting commoditised).
Are they? Why is every model more expensive than the last? Why are company AI spends spiralling out of control?
> It doesn't matter that the margins are high (companies are secretly hiding the real costs because that would pop the bubble).
Then why is everyone except Nvidia losing money?
> In their world view, the bubble is predestined.
You don't need to be a stock market genius to look at a technology that doesn't provide the benefits promised but is being sold as if it's going to change the world and see a crash coming. The bill comes due eventually, and when it does the gap between the actual value and the promised value will need to be addressed. If that's not going to be a "bubble pop" I don't know what is.
The only way the financials work out, the only way there is not some form of bubble, is if advances in AI not only continue at the current rate but speed up, and this technology becomes the most valuable invention humanity has ever made. If you believe that's the case then you're living in a fantasy world so divorced from reality that no evidence I could present would change your mind.
It's widely reported on and discussed. You'll find plenty of people saying it isn't one, but I'm sure you can find an economist willing to take any position you like if you look hard enough. Not to mention everyone who's financially incentivised to continue to claim that there isn't a bubble.
> Even the most skeptical investors to bubbles like Buffet and Burry, have large stakes in Google and Microsoft, front line bets on AI.
Burry has been warning about an AI bubble for years.
> There is essentially nothing magical about being human. Our primitive brains were trained on logic, reasoning and processes. An LLM is simply this on less efficient hardware, but, improving at a speed faster than human evolution.
There is no basis for this claim. There is no evidence that human brains and the process of evolution that led to modern humans is at all similar to how LLMs function. If you're going to evaluate an LLM then evaluate an LLM, you can't say "Humans can do X, LLMs are like humans, therefore LLMs can/will do X" because there is no evidence to connect them.
Let's assume for a moment that you're correct. That AI is already better than most developers, for whatever definition of "better" you like, and they will very soon be better than any person. (I think this is a total fantasy and you've failed to recognise the limitations as the article points out, but I digress.)
In that case the end goal of these companies is to replace all software engineers. Do you not see that? They've not been hiding this fact. It's a good thing for you and I that these models don't work, because if they did the "golden age" is not coming for us, it's coming for people who own compute capacity and the rest of us will become labourers.
The author is factually incorrect here. Moving information out of the model weights and into the input of the model's context window in no way ensures that the model will accurately output content that was input from the context. This is true even when RAG is used to input exactly the correct data.
Nobody is stopping you, please go ahead.
The problem with all of this is that because an LLM gives a facsimile of language and conversation people believe it can "comprehend" or "understand" or whatever word you want to use. This problem isn't new, AI research has had this issue for decades. The risk is that this makes people misunderstand and offload their decision making to it, treating it like a person to be instructed or reasoned with. It isn't, and can never be.
I have met many programmers throughout my career, and very few of them want to talk to stakeholders, much less customers (exceptions are freelancers and founders, especially of software development shops). And, “having clarity on the priorities” boils down to “just tell me what to do and don't switch it up every two days”.
It explained something to me today. So your point is disproven.
What is intelligence then if not explaining things.
LLMs by definition cannot learn, understand, make judgements, or have opinions. If you find the model to be a tool that is useful to you, by all means. But do not anthropomorphize.
[0] https://dictionary.cambridge.org/dictionary/english/intellig...
I suspect there's going to be some good people who leave, not wanting to participate in that sort of dialogue
The Codeberg blog has articles critical of AI published back in September 2024[0]:
If they can use so much power to bring our systems down, imagine how much power they use by doing this to the entire Internet: We are concerned about the environmental and social impact of the current AI hype.
Growth at all costs is something that we, as a non-profit, are not interested in. We believe in frugal computing and that the resources our planet provides are limited. It is no coincidence that we directly "compete" against a company that funnels billions into the AI hype.
[0] https://blog.codeberg.org/letter-from-codeberg-software-is-a...
AI is many things, and a moral choice is one of them. Codeberg's stance isn't a moral one as much as a technical one, but there is a very strong argument that the use of AI is inherently immoral.
I'm faced with the question on whether all my investment can go up in flames on the whims of a shadow group.
There will inevitably be some people who dismiss this story as a dishonest marketing trick by OpenAI to make their models sound terrifyingly effective. I found 81 instances of the term “marketing” in the Hacker News discussion of the incident.
To those people I say pull your heads out of the sand—you’re now including Hugging Face in your conspiracy theories, just so you can deny the crescendo of evidence here!
That's the hard truth: Vibe coders are addicted to using LLMs. You're going to reject that. You're now going straight to typing up a response about how I'm wrong. But addicts often cannot accept they are addicts. It's why the first step is always recognizing you have a problem, because even that is a hard step.
Tech is full of people who think they're smarter than everyone else, which, as always, makes you more susceptible to addiction. If you think "I'm smart, I can't get addicted. I can tell the difference." then no. No you can't.
Writing code helps me think.
If the tools are making you worse, don't use the tools.
The faster you realize this will never pop, the faster you realize that you too can make money in the biggest gold rush in human history.
Gold worth tens of billions of today's US dollars was recovered, which led to great wealth for a few, though many who participated in the California gold rush earned little more than they had started with.
The human and environmental costs of the Gold Rush were substantial. Native Americans, dependent on traditional hunting, gathering and agriculture, became the victims of starvation and disease, as gravel, silt and toxic chemicals from prospecting operations killed fish and destroyed habitats.[0]
Every time.