Can someone explain to me the scope of work a carpenter is expected to carry out inside a data center? Are we just talking reception and the human areas for fit and finish, hanging doors and so on? Sounds pretty limited. Inside the data floor, I cannot recall ever seeing an ounce of wood, anywhere, except for the heads of many of the "intelligent hands".

Only at a very coarse level, to a simple mechanistic grade school understanding, and misses a plethora of the amazing, little understood things neurons, as animal cells, can do, which just goes to show computer scientists aren't neuroscientists.

You seem to getting on your high horse, so a warning: you probably shouldn't base any ideas of being the equivalent of a biological brain, or capable of thought on that alone, skippy. You're fetishising a computer program

LLM generated posts on Slashdot, who'd have thunk it?

Crushing peer pressure and stressing about social status has always been part of being a child. I think a lot of adults forget that, they'd rather listen to the sound of children playing from a distance, without listening too carefully to what they're actually saying. As a Gen X'er, I can say there was always peer pressure and stressing about social status. Sure, before the internet and smart devices took a hold, it was limited to other children in your local area, and you took your aspirational social cues from television and magazines, but it was very, very much there. I don't think older generations had it any different.

When Grok entered its MechaHitler and "white genocide" fixation period, it was an early very clumsy and extremely obvious attempt to shape its ideology to that of its benefactor. There is no surprise at all that models will be tweaked by those with power or influence to output whatever they demand and think is "correct". No doubt it's far more subtle now and you're hardly going to spot it from insane outbursts any more.

"Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them."

Amazing, to think that regurgitation isn't a known problem at all, and that there certaintly aren't several famous lawsuits based on several-paragraph near verbatim output.

Good point, I should have phrased it better. Obviously I am not saying humans don't learn from books! I do not believe LLM training and human learning are in any shape of form analogous. LLMs superficially aim to replicate some of the coarse detail of how biological brains are put together, but the analogy ends there. If you believe they both do learn in the same way, then we are talking at cross purposes.

Take a grain of sand, and you won't see a single sandcastle, truly astonishing. For an LLM, how does training work? To progressively become better at duplicating the input. I've seen the argument that "you cannot see the structure inside the individual points of data, so it cannot exist" before. This must come as a shock to AI researchers, who can identify areas of data associated with specific meanings.

I was responding directly to the claim that some people (with the implication of including me) thought LLMs do things people do not, which was a straw man. As such, I don't believe you understand what "non sequitur" means.

I suspect this was intended as a troll, but it's an interesting reversal. In this case, a few large corporations have taken the sum total of human civilisation (or at least, that part that was available digitally) without compensation, and are selling it back. It cannot have escaped anyones attention that it is the rich who are rent seekers, not the poor. If you are a small creator, you are struggling to create every day just to survive. If someone takes your work, there is little you can do about it, the legal route is too expensive. Meanwhile if you are a corporation, you can take from anywhere you desire, and can expect no blowback unless your target is wealthy enough to fight back. To steal from Harry Brewis, https://www.youtube.com/watch?...

You have it precisely backwards, people do things LLMs don't, not the other way around. I suspect that the proponents of soft AI genuinely believe that LLMs are doing everything that humans do, and that humans merely regurgitate mixtures of what they've been fed. Which doesn't explain the whole upward trajectory of human civilisation. It also doesn't explain why LLMs need real thought, not artificial thought, to train on, or they decohere. Why do most people read books? It's not because we're force-fed them in order to regurgitate them later. It isn't for our survival, there's no biological mechanism compelling us to read.

Well, two points there. First, it's not the same way a human reads a novel. Second, a human doesn't "train" on a novel, because training implies learning to duplicate its structure. The only humans "training" on a novel and proceeding to write their own are what we call cut and dried plagiarists.

Cart before horse. "Evidence to say that left without clocks and relying exclusively on our our internal cycle" should be "Evidence to say that switching from our own internal cycle to clocks" - as the latter is the biological ground state, the former an arbitrary societal constraint. And the evidence is there: https://pmc.ncbi.nlm.nih.gov/a...

This isn't strictly true, if, for example, you sign up for a store loyalty card and as part of sign up terms it gives explicit consent.

"All the smart people were saying it". No, all the rich people were saying it, and all their sycophants were repeating it. A very different kettle of fish.