No, itâs actually a garbled translation of Echse because lizard people.
Donât other industries use cooling towers to recirculate that water?
And it sounds like a lot of waste energy that could possibly be harvested.
And, just limit AI to necessary applications.
Yep, but it seems like Tech, as an industry, has no interesting in how other fields have solved problems, and would rather just âinnovateâ on their own terms - meaning waste all the resources, because THEY ARE THE MOST IMPORTANT INDUSTRY EVER IN THE HISTORY OF INDUSTRYâŚ
Or, just⌠not have it at all. Iâm okay with that.
What? No magic machine that spits out dubious answers a little faster than the ones you can get from just looking it up with whatâs rapidly becoming the old way of looking things up? /s
I know thatâs not the only âuseâ for AI, but it really does seem like the main point is just to (some day) make those who own and control it tons of cash.
I hear about workers here and there finding ways to do various tasks quicker with it, but that seems destined to mean that workers will eventually just be expected to do more work than they were before. All while worrying, in many cases, whether theyâll completely lose their jobs to AI.
I too do not welcome our AI-wielding overlords.
I think at this point there are some questions where AI has started to work noticeably better than regular search engines, not because itâs gotten good but because the search engines are being made so bad at their jobs.
The article didnât say, but I assumed they were talking about cooling towers, and the numbers make sense.
A large chunk of the cooling in a cooling tower is from water evaporating. That evaporated water needs to be topped up.
The water consumption numbers theyâre talking about are consistent with the cooling of a few-hundred-megawatts of heat using cooling towers.
Powerful. Brave. Scary, but also empowering.
This Is What a Digital Coup Looks Like | Carole Cadwalladr
(Sorry if itâs already been posted â it could go in several threads.)
The best use Iâve heard of was a group trying to make a medicine using AI to analyze thousands of chemical compounds and narrow them down to a few dozen that were worth having the human scientists actually test, which led to productive results that would have been impossible to achieve without that aid. Iâm not sure if Iâm remembering every detail correctly, but thatâs the gist- using AI for something itâs actually better at than humans, analyzing and sifting huge piles of data. Not doing the actual research, drawing any conclusions, or making shit up based on what people on the internet posted. I could see AI being a very useful tool for limited applications like that, rather than the more common definition of âshitty lying chatbot shoehorned where itâs not wantedâ that seems popular with tech bros right now.
Iâve used machine learning for classifiers before. It can be a really helpful way to take advantage of a well-built data set. Itâs also nothing new and doesnât need to involve these kinds of data centers, which go with trying to use whatever thoughtless tools to squeeze whatever drops they can from raw data stolen in bulk. They both get called âAIâ but itâs like comparing a ferry for crossing a river to a mega-yacht where any function is an afterthought.
They donât have closed system cooling towers?
Possible, though closed systems canât cool below the ambient dry bulb temperature, and in e.g. Arizona the building interior would need to be cooler than that most days. So they would need a combination system, with the closed circuit carrying only part of the load.
Which really brings us back to âmassive data centres in hot dry places are a bad ideaâ.
âThere is no such thing as a âfishâ.â
General Predictive Text engines are not the same, even a little bit, as an Image Generation AI, or a Machine Intelligence System, or a learning system, or, or, orâŚ
Calling them all âAIâ is a massive bait and switch: âYou canât crack down on ChatGPT! What about this other AI which is curing cancer! Why are you stopping us from building more data centers for GPT? Why do you hate children with cancer?â
Itâs a bullshit rhetorical hack. Itâs like aliens who see all members of the species Homo sapiens as âpersonâ, and you get arguments about how you canât institute laws which imprison murderers because it will punish babies and brain surgeons, and the arguments are all along the lines of âYou canât stop persons, because what about these persons, why do you want to restrict the freedoms of persons when there are persons doing good things?â
I donât want to stop âAIâ, I want to stop generative slop from fulfilling the nanotech fears, and turning our noosphere into grey goo. Expert Systems which are analysing drugs and folding proteins and smoothing traffic flow, theyâre just fine. Small scale chat engines? (Like how Iâve heard of smaller GPT engines trained on specific limited corpuses intended for education, in the form of Socratic dialogue with students) OK, Iâm listening.
But turning the noosphere into grey goo? Fuck all the way off with that shit, and saying that itâs a restriction on âAIâ without specifying further is becoming tantamount to a lie.
For the record, ichthyologists basically use the definition of âa vertebrate that isnât a tetrapodâ and it works perfectly well except for all these online weirdos who somehow got the idea that English words have to be monophyletic groups. They donât even in biology, they just canât be taxa. Itâs really annoying how people seize up on these things and try to be pedantic without understanding them.
I get your point, but this is more like a misleading conflation:
âWhy are you proposing to ban all goldfish?â
âIâm not, I just prefer not to swim amongst hungry sharks.â
âWhy? What have hagfish ever done to you?â
âI didnât say anything about hagfish?â
âAnd yet you propose these restrictions which would stop the good work that lungfish are doing.â
Thatâs what the arguments about AI feel like.
At one level theyâre all basically the same thing. At another, more practical and important level, no theyâre fucking not.