Earlier quoted context omitted.
They were not asking for your individual opinion. They were asking why any given person would need it.
It turns out I am any given person
You are missing the forest for the trees.
AI is slowing down
791–800 of 820 posts
Re: AI is slowing down
#792Before you spend 20 minutes reading this article, it's worth understanding that the writer has been posting popular but consistently wrong takes for 2+ years (e.g. https://www.wheresyoured.at/peakai/ from March 2024) arguing that AI is failing, is a waste of money, is bad, will never work, etc.
[flagged]
Please don't comment like this on HN. We can disagree with people whilst still affording people basic decency; it's a condition of participating here.
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
Re: AI is slowing down
#793Zitron is begging for a collapse at this point. Yes, his macro analysis correctly identifies a massive financial risk but his incessant pessimism completely misses the incredible ground-level utility that many of us on HN celebrate every day through undeniable, massive productivity gains. At this point I'm trying to believe there's a middle ground where the level of individual capability this unlocks, leads to major…
[flagged]
Please don't comment like this on HN. We can disagree with people whilst still affording people basic decency; it's a condition of participating here.
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
Re: AI is slowing down
#794Earlier quoted context omitted.
> He is not writing his blog to convince people, his primary audience already agrees. He's selling a paid newsletter, so at least one of his motivations is to make money. His target subscribers are certainly people who lean towards his viewpoint but he still needs to do some convincing because the market of people who are open if not warm to his thesis is much bigger than the market of people who already share his th…
Subscription based writing is all about writing for audience that agrees. Yes it creates bubbles, but economic of it is "people paying to read stuff they agree with".
I think this author makes a number of good points. So I'm part of that audience "that agrees" to a large extent.
But I would never consider subscribing to this newsletter and the biggest reason is the writing style/tone. I find it unpleasant.
So even when people agree with you, you can lose them by being abrasive, unpleasant, unlikable.
Re: AI is slowing down
#795* * *
>LLM-type AI exacts huge costs because it is terrible at reporting "I don't know". When it doesn't know, it generates noise and polishes it. >If a "confidence too low for output" signal could be extracted, this whole technology would be a lot more useful Anthropic's interpretability research explored this topic a bit in 2025. Apparently, the signal is extractable: It turns out that, in Claude, refusal to answer is th…
That's real progress. The paper behind it is [1] They try to extract "attribution graphs" to understand why the LLM produced some result. It's encouraging to see more work on what's going on inside. They obtained insight into a specific kind of hallucination: not finding a specific fact. "We uncover circuit mechanisms that allow the model to distinguish between familiar and unfamiliar entities, which determine whether it elects to answer a factual question or profess ignorance. “Misfires” of this circuit can cause hallucinations." That should be tested on queries which resulted in making up legal citations.
[1] https://transformer-circuits.pub/2025/attribution-graphs/met...
Re: AI is slowing down
#796Earlier quoted context omitted.
There is no substance to the article to tackle. Ed Zitron is a clown that makes seven million dollars a year writing unsubstantiated bullshit on the internet. He has been consistently wrong in every single one of his predictions and will continue predicting the sky is falling, because his content has zero educational value. It is purely a confirmation-bias inducing dopamine hit for the perpetually aggrieved.
Hypothetically speaking, how does one go about being so handsomely rewarded for being wrong about everything? I feel like I am well qualified for this type of job.
You never know if people who are running a skeptic grift are at the same time long the very trends they're criticizing.
I agree with the general thesis that AI is a bubble and there's a lot of over/mal-investment, but that hasn't stopped me from riding the wave and making a considerable amount of money investing in AI-related stocks.
If I was more ambitious, I'd write a $70/year newsletter ranting about how AI is a massive bubble that could collapse at any moment to make money on both sides.
Re: AI is slowing down
#797How are the author's missed predictions different from the missed predictions of "AI" believers
For example,
Re: AI is slowing down
#798Earlier quoted context omitted.
If I thought there were some actual small cabal of people running the global economy, this is almost like a novel: massive amounts of money entered the economy starting in 2008 and 2020-2023, the rich became insanely wealthy. Their wealth is now all tied up in the 2020s version of the railroad/fiber, we're going to essentially erase trillions of dollars from the global economy and reset. We sure do need a reset.
I really doubt the US will erase any money any time soon. The reset is prone to happen by other means.
Rapid inflation/devaluation is certainly a way to do it.
Re: AI is slowing down
#799Earlier quoted context omitted.
I particularly enjoy reading big banners asking me to pay for a newsletter subscription if I "liked" the content. Not if I found it interesting. Not if it actually provided any value whatsoever to me. No, you just have to "like" it. In other words, it is meant to be written in an engaging way and perhaps reinforce your believes like an echo chamber or even stir up certain strong emotions. Not to convey information. S…
What are you talking about? Why is "liking" something mutually exclusive with conveying information? I like lots of things precisely because they convey information!
>I like lots of things precisely because they convey information!
Correction: You may like them because you think they convey information. But without any sort of vetting process, the internet has become a cesspool of "news" or "general knowledge" places that ended up quite successful, but which are essentially just a contest of who is most confident when talking about topics and who can present bullshit in the most engaging way. You can see the peak of this on the JRE podcast. Anyone with actual expertise in a subject would be able to call out many of the guests, but since the host knows nothing about most of their fields he just gives them a platform to spread their opinions as facts. And millions of people who also don't know better will accept them without question.
Re: AI is slowing down
#800Earlier quoted context omitted.
Good questions. The post you are replying to is also based on vibes. So how about we make a bet and come back after IPO filings? But isn't it okay to suggest that random 3rd party hosting companies are ... not losing money? Why?
I'll bet 50€ that big AI companies (OpenAI, Anthropic, xAI) have bad regarding their core financials. Excluding one time deals such as renting hardware, external cash infusions, the core AI model business (training models and selling inference) is unprofitable and will be be at the expected scales (50 billion € or more) for the next 5 years at least. I bet that within 5 years they will be sold for scrap to bigger com…
This is not fair, lets just speak about Inference margins. You are bringing 100 other things to make it more ambgigous.
Statement: margins API prices of all models are greater than 10% in Anthropic.