From what I've seen lately, Ed isn't saying the technology isn't significant. He criticizes the fact that AI companies are dumping this kind of money into it with circular financing deals. If it goes like you say, (GPUs at home) the AI labs will be destroyed, because no one will need the inference capacity they're dumping money into when they can do the decent part at home. From what I've seen he says it's more "not worth the investment" rather than not a breakthrough. It just sucks compared to the trillions they're spending which they'll never recover, is the argument. But I guess it's easier to strawman the argument to "this stuff isn't going to change anything".
He is saying technology isnt significant. Again and again and he puts a lot of emphasis on it. He is also focusing the circular financing deals, on it being a bubble etc.
Obviously, you can agree with one thing and not with other. And he can be right in one thing and not the other. But, if you listen to what he is saying, he is absolutely saying the technology is not significant.
I think you're misinterpreting a strongly worded "it isn't as great as the hype" with "there's nothing there". Because "nothing there that matches the hype" is both 100% true and completely different than "nothing there".
I don't think I am misinterpreting him. I did read some of his articles and did listened to some of his podcasts.
He is very very clear and open on what he thinks about usefulness of ai. He is not saying "it isn't as great as the hype". He is saying "it is useless". He is simply not the centrist kind of guy when it comes to AI usefulness.
Indeed it's worse, and apparently Valve/Steam is the only one who seems to care about something resembling freedom to sell legal things, even if we might subjectively disagree.
Well, this definitely got pushed down in the HN ranks. 49 points in 1 hour, but let's not talk about anything that might cause people to rethink technofascist human replacement.
If that's claptrap this is article sewage waste. On the nose is an understatement.
It's not doing the same work, or producing the same output.
It's an episodic response to specific requests. If you went into a coma between every question you were asked, you wouldnt be in a normal human state of consciousness either.
This article pretends that we understand human brains as much as we understand the simple algorithms of LLMs. And that's just laughable. Even so out of touch as to say consumption of "[humans] consume thousands of liters of water per years". As if there isn't multiple orders of magnitude more consumption for data centers. For data centers that produce a braindead simulacrum.
>If you went into a coma between every question you were asked, you wouldnt be in a normal human state of consciousness either.
I don't think any one is saying AI has a normal human state of consciousness.
Nobody claims that people with anterograde amnesia are not conscious. Similarly, you could drug someone unconscious after every time they answer a single question. It's not a normal state of human consciousness, but I wouldn't want to say a person in that situation were not sentient.
It is true that there is much that we do not know about the human brain, but it is also true that many of the things that people describe as attributes that clearly disqualify AI have known analogues in humans.
From simple failures of perception in optical illusions, to the inability to notice quite significant changes in front of your eyes if the change happens during a saccade. There are a wide range of known limitations that humans have that reveal how much we overlook things we can't do, that when you consider them, reveals our own behaviour could be an more of an arrangement of systems than we would like to think.
Every single one of your examples is one that comes nowhere near the blankness of an llm pile of numbers when it isn't answering a question. Maybe I should have said, is completely dead, rather than in a coma, between questions.
The quirks we observe in llms we only pretend have analogues in humans because we love to anthropomorphize. And it's easy to ansthropomorphise a language simulacrum. We've been doing it since ELIZA.
I'm not sure why you think a temporary cessation of activity is relevant at all.
Would you think of a hypothetical machine that could freeze a human to absolute zero and restore them again sufficient to declare their behaviour when unfrozen as not that of a sentient individual.
How about if you had a picture of a black man and a white woman in front of a person and you swapped the heads, right in front of the observers eyes, but you picked just the right time so that even though they never sensed any time of not looking at the image, they did not notice the change.
This is an example of how llm's break when you ask them "what should I build?" No current AI can provide good judgement.
To be clear: gaussian splat is very cool, but "in the terminal" is just nerdbait. The project includes some explanations, but given the llm source, I'm not sure it's accurate.
Here is a good video describing the tech in general, if you're wondering why there's so much buzz about it:
We were talking about whether these metrics are meaningful. I was just pointing out that even a tiny one-person company can burn a lot of tokens.
As to whether the token spend is questionable, the number I quoted is for my production AI pipelines, not for coding. And my customers (and profit margin) seem to think the spending is valuable.
Questionable spending aside, GGP is providing information about how a specific metric may not measure what people think it measures. There is value in that comment.
I'm old enough to remember when conservatives thought the government picking winners and losers in industry was a bad thing. But that was before about 75% of them joined the dumpty cult.
reply