Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Demis Hassabis has a plan to harness AI safely

twitter.com

Post reply on HN
Print view View original

81–90 of 217 posts

Re: Demis Hassabis has a plan to harness AI safely

#81

Post by verdverm »

estearum wrote: post #38
thegrim33 wrote: post #10

Spoiler: The plan is .. add massive regulation, but only to the US, don't affect other countries developing it in any way other than "setting a good standard that'll hopefully influence them". Seems like an airtight plan.

That's called the carrot The beauty of the United States' global hegemony is that it also has lots of sticks

Those sticks look less effective today. The hegemony is in recession as the rest of the world smiles, pays omage to dear leader as needed, and works to decouple in the background.

Re: Demis Hassabis has a plan to harness AI safely

#82

Post by verdverm »

techpression wrote: post #42

Has anything really important been solved by AI yet, or where is this radical (imo) belief around AGI coming from? Genuinely curious, I know there are some math problems solved and ML has been used for far longer than AI to improve things, but where is clean (efficient) energy, the cure for cancer (or any of the horrible neurological disorders, take your pick), new hardware designs, quantum computing solutions, etc e…

AlphaFold is a pretty big deal, especially the millions of protein folds they published openly and scientists are just digging into.

Re: Demis Hassabis has a plan to harness AI safely

#83

Post by knocte »

tehlike wrote: post #78
knocte wrote: post #76

Am I the only one thinking this tweet is just a word salad with a lot of sauces and condiments?

Word salad for regulatory capture.

Well, if he had proposed more concrete plans for what that body might do... but yeah, it seems too abstract that indeed sounds like a bureaucrat wanting to wrap everything under a big pile of paperwork.

Re: Demis Hassabis has a plan to harness AI safely

#84

Post by SirHackalot »

lompad wrote: post #77
squidbeak wrote: post #33

Earlier quoted context omitted.

If the USA takes up his suggestions, it will have an incentive to work on international frameworks and treaties with competing nations to make the regulation global.

There won't be any kind of international framework, because by now countries have learned that the US' word is pretty much worthless, similar to Russia's. They now know from experience, US-ratified treaties will only be honored if they feel like it on a particular day. If it's inconvenient, it will simply be ignored. You'd have major protests in most large economies if they deliberately put themselves under the boot…

Wish you were wrong… The trust in U.S. as the world’s policeman has eroded dramatically.

Re: Demis Hassabis has a plan to harness AI safely

#85

Post by lmf4lol »

minraws wrote: post #79

Demis has only 1 plan, how to dodge releasing a new model at this point, jokes aside I value thinking about AI safety, but are we really so close to AGI? It doesn't feel like it, LLMs still diagnose my headache as a chronic illness or a brain tumor from time to time... honestly stressful.

regarding your headache diagnosis. If you just ask: could my headache be a brain tumor, the answer is probably: very unlikely but it might be. thats because its actually true. there is actually a non 0 chance that headache is a symptom of a brain tumor. From a bayesian standpoint though its really super super unlikely, especially if there are no other symptoms that might shift the prior.

a better way to use an LLM here is to let it research scientific papers (or medical guidelines) on headaches. then give it as much info on your symptoms as possible, and then ask it do deduct potential diagnoses. You can even ask it to calculate probabilities if thats what calms you down. Probabilities based on studies and available information. IMHO, this always leads to a more rational response.

maybe you are doing this already, idk. just wanted to share what works for me.

Re: Demis Hassabis has a plan to harness AI safely

#86

Post by pshirshov »

estearum wrote: post #34
pshirshov wrote: post #19

Blah-blah-singularity, so let's cripple the models so much they refuse to talk about React, because who knows if you are not cooking chemical weapons or meth in your browser's DOM, right?

strawmen are fun and helpful

Fable literally refused me to chat about React several times citing limitations on chemical weapons development.

Re: Demis Hassabis has a plan to harness AI safely

#87

Post by ignoramous »

merelydev wrote: post #46

How will this be enforced, at least with financial markets money is discrete, can largely be counted. This seems like a slippery slope to full blown surveillance of the internet and in general computing. If AGI is truly imminent and will collectively effect all of us why not apply democracy to it, and vote for new AI models?

> will collectively effect all of us, why not apply democracy to it

That's because most execs proposing solutions are "Technocrats" or think like one: https://en.wikipedia.org/wiki/Technocracy

Besides, I don't think as a collective we're well equipped to decide one way or the other. If the collective were given a say, billions will be spent, often in consultation with technocrats, on doom / hype marketing (if it isn't happening already).

Re: Demis Hassabis has a plan to harness AI safely

#88

Post by hackinthebochs »

mikeyouse wrote: post #58
noelwelsh wrote: post #3

The premise is "Artificial General Intelligence (AGI), a system that exhibits all the cognitive capabilities the brain has, is probably only a few short years away". If this is true, establishing an institution to ensure things like "publishing model cards with technical details, maintaining strong internal cybersecurity, vetting key personnel, and providing sufficient resourcing for safety and security research" is…

My great annoyance with all this blather about UBI and post-scarcity and new economic models is that it relies on taxing and redistributing the ~infinite profits of the largest tech companies on earth. What in the history of our world gives anyone faith that those companies are going to start paying taxes instead of using "AGI" to engineer increasingly complex methods to avoid them so that their equity owners can poc…

As long as the proles have a monopoly on force (through their representative government), the AI companies will pay taxes. The question is how long does this state of affairs last.

Re: Demis Hassabis has a plan to harness AI safely

#89

Post by modeless »

This genre of AI researchers begging third parties to stop them from destroying the world is getting really tiresome. Setting up regulatory bodies with ill-defined goals to counter hypothetical threats from science fiction is a recipe for disaster.

The premise is that government is too slow moving to be able to react once actual problems are discovered, and I reject that. Yes, government usually moves slowly in most circumstances, but given singular existential threats it definitely can move quickly. Instead of acting out randomly and tying ourselves down with speculative regulation that probably won't even address the real problems, we should wait until the problems are obvious and then act decisively with targeted fixes.

Re: Demis Hassabis has a plan to harness AI safely

#90

Post by techpression »

SpicyLemonZest wrote: post #50
techpression wrote: post #42

Has anything really important been solved by AI yet, or where is this radical (imo) belief around AGI coming from? Genuinely curious, I know there are some math problems solved and ML has been used for far longer than AI to improve things, but where is clean (efficient) energy, the cure for cancer (or any of the horrible neurological disorders, take your pick), new hardware designs, quantum computing solutions, etc e…

The point is that we need to have a safety plan in place before an AI is smart enough to radically reshape the world. If you've got an AI that's ready to start sending humanity into the next era of civilization, it may be too late to control when and how it does that. > Heck, a proof for P=NP or P!=NP or solve the The Riemann Hypothesis. Just give me something truly exciting and I will believe AGI is around the corne…

Why would we be able to design a safety plan that would control such a powerful force? It would be like putting umbrellas on an asteroid hoping to slow its fall. That sounds like delusions of grandeur.
Post reply on HN