The UK's "AI Opportunities Action Plan" (https://www.gov.uk/government/publications/ai-opportunities-action-plan/ai-opportunities-action-plan) perpetuates the fallacy that we should embrace the "AI revolution" and its huge increase in demand for compute to make society better through e.g. better diagnostic tools in healthcare, smarter electricity grids, etc.
The growth in "AI" is entirely driven by generative AI based on Large Language Models. It has little or nothing to do with the Machine Learning approaches that already help with so many aspects of society. (1/2)
=> More informations about this toot | More toots from wim_v12e@scholar.social
Compared to GPT-4 or Gemini, these are tiny models that do not require huge amounts of compute, memory or storage. Where GPT-4 is more than a trillion parameters (10^12), a typical CNN for medical applications (*) is of the order of tens of millions of parameters (10^7). So it needs a hundred thousand times fewer resources. And many of the ML techniques that do not use neural networks, like SVM or random forest, are even smaller than that. (2/2)
[#]FrugalComputing
=> More informations about this toot | More toots from wim_v12e@scholar.social
(*) For example, SegNet, a network shown to have 99% accuracy for colon cancer detection, is 7.6M parameters (https://www.sciencedirect.com/science/article/pii/S0010482521005242). An LLM with 7B parameters (1000x more than that) can run easily on a laptop (https://www.hardware-corner.net/llm-database/LLaMA/).
=> More informations about this toot | More toots from wim_v12e@scholar.social
@wim_v12e That's quite interesting. A bit of the problem that we've got now is that LLM and AI has somehow become synonyms in the larger discourse, drowning some of the finer points in tech-bro-babble.
=> More informations about this toot | More toots from zerodogg@fosstodon.org
@zerodogg @wim_v12e I guess it's easier for the tech industry to sell "intelligence" than "models" or "deep learning"?
=> More informations about this toot | More toots from trobador@mastodon.social
@trobador @wim_v12e That's certainly one side of it, but it does muddy the waters, since LLM and most of gen-AI is a garbage fire, while these models, used correctly, can actually help. With emphasis on "used correctly".
=> More informations about this toot | More toots from zerodogg@fosstodon.org
@zerodogg @wim_v12e I agree, and I believe this conflation is intentional: by grouping LLM chatbots with everything related to machine learning these big tech spokespersons and lobbyists can say things like "AI is helping cure deadly diseases" and refer to the critics as "irrational luddites"...
=> More informations about this toot | More toots from trobador@mastodon.social
@trobador @zerodogg
It's definitely intentional from the side of OpenAI and the likes.
But for the UK government to perpetuate the same notion is really disappointing.
=> More informations about this toot | More toots from wim_v12e@scholar.social
@wim_v12e @trobador It really is, but they're in good company - there's a lot of "we need to use AI" around, where the using is more important than actually solving any real problem.
=> More informations about this toot | More toots from zerodogg@fosstodon.org
@zerodogg That is a big part of it. In the public perception, it's all "AI". It's why I always use ML rather than AI (and also because there is nothing intelligent about it).
=> More informations about this toot | More toots from wim_v12e@scholar.social
@wim_v12e
I remember a quote from a lecturer at Queens University (I think), that calling the field "AI" is like calling medicine "healing": while AI is an eventual goal, it's not the current situation. Of course that quote is 20 years old, so may be a little out of date...
@zerodogg
=> More informations about this toot | More toots from ddlyh@topspicy.social
@wim_v12e Even the smallest amount of technical literacy from the people asking tax payers to spend billions on this would be nice.
It’s staggering how much Labour have bought in to the ‘LLMs are basically sentient magic’ con, and the arms race framing that benefits the tech investor/planet destroyer class.
=> More informations about this toot | More toots from True_Heresy@mastodon.social
@wim_v12e
When you see the super enthusiastic statements from Microsoft, Amazon, OpenAI, Salesforce and Oracle, you can be confident they stand to gain WAY more than the UK public.
Which is probably why the PM made sure to include the statement about "more money in the pockets of working people" in his press release. In at least four different places.
=> More informations about this toot | More toots from airwhale@beige.party
@airwhale Would be nice if saying it made it happen.
But what bothers me is that, surely, the cabinet has advisors that should tell them that the probability of this plan creating economic growth is vanishingly small? It's so transparent.
=> More informations about this toot | More toots from wim_v12e@scholar.social
@wim_v12e
I've got plenty of smart colleagues who are super enthusiastic about AI tech, and the scary thing is they seem to ignore or accept all the risks. Management is singing the same song too.
Seems to me more like blockchain, crypto and Meta-verse all over again, only this time, even more agressively pushed by tech giants who start to realize they probably oversold their product. They've sunk SO much money into this already so they MUST produce returns.
Of course, we /could/ be wrong, but I've got a bad feeling about this.
=> More informations about this toot | More toots from airwhale@beige.party This content has been proxied by September (ba2dc).Proxy Information
text/gemini