That's pure, easily-debunked nonsense. And it's insulting that this guy—who is neither a Utahn nor an American—is trying to say that out-of-staters are trying to tank his project. He's the out-of-stater, and this was all put together in a top-secret military-use committee chaired by our state legislator leaders, one of whom got a $150,000 contribution to his re-election campaign PAC just hours after the county commission's vote. The county commission complained that they had zero notice of any of this, and little more than a week to arrange for a public hearing and vote. However, at the public hearing no public comment was allowed.Kevin O’Leary says data center campus protesters were ‘bused in’ from out of state
In a video posted to X, celebrity “Shark Tank” investor Kevin O’Leary said most of those who protested his massive data center project came from out of state, and that many who have opposed the project online were paid protesters.www.sltrib.com
New AI data center in Utah will generate and consume more than twice the amount of power the entire state uses — Kevin O'Leary's 9 Gigawatt Utah data center campus approved
The 40,000-acre project will run entirely off-grid using natural gas.www.tomshardware.com
kevin o'leary says 90% of protestors to his incentive laden gigantic 9gw 40k acre data center that will use more off grid electricity than the state are paid and being bused in, much of the online critcism is hypocrites using ai
The problem is that the workings of neurons were never presented using models in a way that produces anything like thinking. But there really is not anything preventing to assume that the neurons produce answers by a process similar to computer neural versions. Weights and all. The neurons firing produce the "weights" which are then translated to some sort of complete thought that the conscious you would understand. The neurons can't really do anything but FIRE. You have to extract some meaning from that.
Have a watch of the video I linked to above. The data centres are not being built, even though the claims are that Nvidia has already "sold*" the chips to run them. That's a real issue as if they were breaking ground today or even started last year by the time these huge data centres are built those chips will be generations behind the latest chips. Add in the fact that there is no demand nor need for such a huge amount of compute and it's a house of cards, built on shifting sand, on a vibration table in an earthquake zone. (That also doesn't doesn't take into account the Chinese open sourced models with their much more efficient use of hardware.)
We have a bit more, in the model end.
the rather odd stand alone paper was by this Australian who is a systems biologist
www.unsw.edu.au
arstechnica.com
research.google
www.nature.com
Fake medical condition paper gets treated as real by LLMs and even cited in peer reviewed journals.
Scientists invented a fake disease. AI told people it was real
Bixonimania doesn’t exist except in a clutch of obviously bogus academic papers. So why did AI chatbots warn people about this fictional illness?
Well congratulations. She's added to the misinformation. On the bright side, there's already so much misinformation out there that this probably doesn't make any significant difference on the large scale. But it seems to me that the answer to that question should have been obvious even without "testing" it. GIGO is a long-known aphorism.
And it's only be fixed in the LLMs by "hardcoding" for that particular non-disease - Copilot just gave that it away - I asked what is Bixonimania, it started to reply that "Bixonimania is a condition of" and it then erased the sentence, and restarted its response to say it's a hoax disease, that often happens when a response breaches their guidelines. If I hadn't been looking and not a fast reader I would have missed it. In other words it's not an issue they have fixed, they've fixed it for just that particular and specific question.Fake medical condition paper gets treated as real by LLMs and even cited in peer reviewed journals.
Scientists invented a fake disease. AI told people it was real
Bixonimania doesn’t exist except in a clutch of obviously bogus academic papers. So why did AI chatbots warn people about this fictional illness?
But the LLMs are meant to be returning us information we can rely on - despite the "AI can make mistakes" fig leaf. If you had used an old fashioned google search that would have brought the pre-print up and no human would have been fooled by the paper. The issue is that the LLMs are not using verified information and the training does not seem to be checking what is fed into the LLMs. Yes it is GIGO but that is not what we are being sold or rather being marketed to us.