Yep, we’ve all been saying it, now they’re finally admitting it. LLMs are not actually intelligent and this method is not sustainable in its growth. Finally seeing diminishing returns; thank goodness we can all move onto the next grift now https://sigmoid.social/@garymarcus/113455684170107190
American friends, I honestly have no words. I am so devastated for you, us, the planet, everything right now. I can’t even find it in me to be angry, I’m just so sad for us all
This is the most horrifyingly unethical research I have seen in a long time. How on earth is the University of Wisconsin's IRB allowing this sort of thing to happen? Getting all of this data without consent from students at a university, even working out who their minor-age friends are?
Ah yes “respect for the academic tradition”, that’s what we should be prioritising over EDI, more like “respect for the backward opinions of old white men in academia for whom the advent of EDI is the first time they’ve ever been challenged on their prejudices”.
Richard Stallman is a repulsive human being and I honestly don't understand why the FSF & community has enabled him for so long. (I recognise the excellent work of several in the community who have stood up to him but he seems to be very difficult to dislodge, and that's because he still has enablers.)
https://stallman-report.org/ (cw discussions of defences of child sexual assault, sexual harassment, and sexual assault)
I was quite active in the FS/OS community back at the turn of the century but ended up getting out because these creepy vibes were pretty much everywhere at the time. You just sort of put up with it in the moment, but it really did become "death by a thousand cuts" - constantly being hit on, constantly having to deal with graphic, "joking" discussions of sex or pornography, being undermined because how would women know anything about tech, etc. It got very very tiring.
This is apparently doing some numbers so I want to make it clear that I think this paper is purely part of a strategic move by Apple to save face and gracefully exit from the genAI hype cycle. Hopefully other companies will do the same.
No shit “this decline [in the capability of LLMs to perform logical reasoning through multiple clauses] is due to the fact that current LLMs are not capable of genuine logical reasoning; instead, they attempt to replicate the reasoning steps observed in their training data.”
Paper from Apple engineers showing that genAI models don’t actually understand what they read but just regurgitate what they’ve seen like a puppy wanting to please its owner: https://arxiv.org/pdf/2410.05229
I'm making a scarecrow for our scarecrow festival and what I would really like to do is have a button that a passerby can press and then a song plays. So I'd need weatherproof speaker and remote/button/something, and something that would just play one song. Is there anything off the shelf that would do this? Looking at my #emfcamp peeps because this would probably be up their alley. (bonus points if I can attach fairy or disco lights to the button that come on too)
Going to vote tactically. I wish I didn’t have to as I really am not on board with the Labour Party right now but as you can see my vote against the Tories actually matters here. #ukpolitics
Investigating Matrix to replace google chat, does it really matter what server you're on? Are there any specific benefits to rolling our own? (or is it just a potential vulnerability?) We'll be using it for e2ee. #matrix
Professor of Ethics and Games Technology, University of Staffordshire. Co-I IGGI CTD. Vice Chair ACM Committee on Professional Ethics. Interim EiC ACM Games: Research & Practice. Responsible innovation, tech ethics, video games, “AI” & crypto sceptic. Erdos: 3; Bacon: 3. She/her. Moderator mastodon.me.uk, content co-lead for #EMF2024. Naturalised British, Aussie immigrant from Gamaragal country, mum to 2 humans and 4 chickens. Rural Leicestershire. Dances Morris with Black Annis.