4 ms·
"Please stop training these AI" is almost comical, even if they are right and the signers are significant. It would be far more effective to lobby for legislat
by brucethemoose2 4y ago
"Please stop training these AI" is almost comical, even if they are right and the signers are significant.
It would be far more effective to lobby for legislation on, say, liability related to the use of AI, data privacy (especially when it comes to training datasets), psycological manipulation, actual AGI research bounds and so on. They are just burning donator money with messages like that, and the research is not going to be bound without law or incentives.
- jakeinspace 4y agoSure, or start covertly threatening the lives of AI researchers, if you’re that convinced it’s going to end humanity.
- brucethemoose2 4y agoTBH that would just give Big Tech an edge since their research is more "underground."
- contravariant 4y agoIs anyone even trying to develop an AGI by the way? I mean a language model is neat, may even pick up some emergent features, but in the end it has no real identity. It's basically a memoryless computer doing improv where it tries to pick continue a conversation someone else had.
- brucethemoose2 4y agoMaybe not a "real" AGI, but a system that could adequately impersonate an AGI with improv and self training seems quite practical in the near future. https://en.m.wikipedia.org/wiki/Philosophical_zombie https://en.m.wikipedia.org/wiki/Philosophical_zombie
- MacsHeadroom 4y agoHumans are my favorite kind of philosophical zombie. Not a single one has ever given a sound argument that they are anything else.
- quonn 4y agoSo where is the subjective perspective of looking into the world coming from? „It emerges“ is doing a lot work and explains nothing.
- MacsHeadroom 4y agoWhat subjective perspective? Show it to me. Silly pzombie, you can't fool me!
- versteegen 4y agoIt used to be that only a few academics were actually working on AGI, and even they had to pretend they were working on narrow AI to get funding. But now, plenty of big labs are. For example, DeepMind's About page states: > Our long term aim is to solve intelligence, developing more general and capable problem-solving systems, known as artificial general intelligence (AGI). In the long term, I still think DeepMind will get there before OpenAI, due to the direction of their research program, which has always been about embodied agents. I have heard Shane Legg say that embodiment is probably necessary for AGI.
- robwwilliams 4y agoAgreed. DeepMind and Hassabis are on the right track—-a society of modules with many layers of recursive self attentional gating. LMM is one big part though.
- larve 4y agoThe message is "pause training for a few months so we can figure out how to approach liability, safety and regulation". You are fighting a strawman.
- brucethemoose2 4y agoFair, thats really what I meant to type, but its still an impractical demand. Most entities pursuing this aren't going to stop training because some high profile people asked them nicely.
- reasonableklout 4y agoI don't see petitions like this as demands, but as coordination mechanisms. If there are only a few industry actors with the resources to train gigantic AI models and they each want to stop but won't because they don't know if their competitors will, it's good for them all to have 1 place to say "I want to stop". Alternatively, if there are tons of industry actors racing and the only way to stop or slow down training is via governmental action - and the government won't act because they don't know how it will affect the industry - it's good to be able to point to "that one petition with 1000 highly influential signatories from the industry asking for regulation and explaining its effects".
- larve 4y agoI think the value in these kind of letters is about spreading a message, building a community and allowing people leaving their signature to "formalize" their position, in a sense. It doesn't cost very much to do it but has a wide reach.