Pause Giant AI Experiments: An Open Letter • Hacker News Board

Re: Pause Giant AI Experiments: An Open Letter

#671

Post by jamilton » Thu, Mar 30, 2023, 1:11 AM UTC

lannisterstark wrote: post #405

reso wrote: post #380

To those saying "this is impossible in our society", there is a long history of scientific fields mutually agreeing to pause certain work until safety and regulations could be put in place. The Asilomar Conference is an example. [1] The idea that "you can't put the genie back in the bottle" is very popular in tech culture, but many have suggested that simply extending elements of copyright or tort law to AI would be…

It's not about it being impossible. I guess the major issue right now is that people are focusing on the wrong aspect of it. They should be focusing on how most LLMs are currently only controlled by bigger coroporations. Democratizing the LLMs and Giving access to people would resolve the x-poly we have currently. This also solves the "safety and regulations" issue to an extent - Do you really trust that Microsoft, G…

What do you mean by "democratizing the LLMs"? There are already open source LLMs. Making Claude, GPT-4, and whatever else is being worked on open source wouldn't make them any less capable of doing harm.

Re: Pause Giant AI Experiments: An Open Letter

#672

Post by Ataraxy » Thu, Mar 30, 2023, 1:11 AM UTC

This is meaningless.

"Good guys follow rules, bad guys don't."

As if getting the infrastructure to privately build your own model is challenging for people that have the capital to spend on such projects.

Re: Pause Giant AI Experiments: An Open Letter

#673

Post by anoy8888 » Thu, Mar 30, 2023, 1:12 AM UTC

xianshou wrote: post #355

Two-party iterated prisoner's dilemma is hard enough. Sensible players will coordinate with something like tit-for-tat, but that only works when both parties start off on the right foot. Regardless of initial strategy, the chances of degenerating towards the mutual-defection Nash equilibrium increase with the number of parties. The only prior example of world coordination at this level would be nuclear disarmament ac…

The real upside of this is that it may get attention of lawmakers and our politicians. However, it is probably not even in their top 100 list of priorities

Re: Pause Giant AI Experiments: An Open Letter

#674

Post by kiney » Thu, Mar 30, 2023, 1:13 AM UTC

alfiedotwtf wrote: post #651

xianshou wrote: post #355

Two-party iterated prisoner's dilemma is hard enough. Sensible players will coordinate with something like tit-for-tat, but that only works when both parties start off on the right foot. Regardless of initial strategy, the chances of degenerating towards the mutual-defection Nash equilibrium increase with the number of parties. The only prior example of world coordination at this level would be nuclear disarmament ac…

This is the first time I've actually seen game theory applied to an actual real life scenario rather than just a theoretical academic problem. What's the best resource to learn more game theory and to be able to applying it like you to a more general setting?

If you happen to understand german I can really recommend "ProfRieck" on youtube

Re: Pause Giant AI Experiments: An Open Letter

#675

Post by jamilton » Thu, Mar 30, 2023, 1:17 AM UTC

dudeinhawaii wrote: post #646

hn_throwaway_99 wrote: post #413

Earlier quoted context omitted.

"Virtue signalling"? Please. There are a lot of very smart experts on that signatory list who definitely don't need to, or care about, virtue signalling. Fine, ignore Musk's signature, but I don't think luminaries like Stuart Russell, Steve Wozniak, Jaan Tallinn or John Hopfield are doing this for "virtue signalling". You can fairly argue that this will be ineffective, but a lot of experts in this field have real, "h…

Where the pause for self-driving cars? How many people have died from that relentless push versus ChatGPT? Very convenient and at the same time silly.

419 accidents involving self-driving (level 2 and 3), 18 deaths, 19 accidents with injury level unknown [0]. All deaths from level 2 vehicles. So being pessimistic, maybe 50 deaths from self-driving.

The people signing this are worried about AI that doesn't exist yet. No one died from nuclear weapons before they were invented.

[0]: https://www.slashgear.com/1202594/how-many-people-have-actua...

Re: Pause Giant AI Experiments: An Open Letter

#676

Post by hdudhdbsbs » Thu, Mar 30, 2023, 1:24 AM UTC

musicale wrote: post #76

petilon wrote: post #68

> Should we automate away all the jobs, including the fulfilling ones? You don't necessarily need jobs because you don't need money. What you need is wealth. I am going to leave it to PG to explain the difference [1]: Wealth is not money. Wealth is stuff we want: food, clothes, houses, cars, gadgets, travel to interesting places, and so on. You can have wealth without having money. If you had a magic machine that cou…

> AI & robots will generate wealth at unprecedented scale, and humans will reap the benefits, the same way Alaskans reap the benefits of their natural resources. Vast wealth will certainly be delivered to owners of the robots and AI systems. What reason is there to believe that they would share the wealth with ordinary citizens? A typical alternative to the Alaska model is one where oil companies extract all of the w…

We just need to vote in a new rule for a national dividend to be paid out to everyone. Problem solved

Re: Pause Giant AI Experiments: An Open Letter

#677

Post by uptownfunk » Thu, Mar 30, 2023, 1:26 AM UTC

This is like thrashing on engineering design instead of implementing and shipping. This isn’t something we’re going to figure out through intellectual debate, the only way it will happen is if we make it, bad things happen, and we learn from it. As much as I wish it weren’t true there are too many degrees of freedom and flexibility that we won’t be able to anticipate. It will be messy but I think messy is the only way through at the moment.

Re: Pause Giant AI Experiments: An Open Letter

#678

Post by jamilton » Thu, Mar 30, 2023, 1:29 AM UTC

Barrin92 wrote: post #554

biscottigelato wrote: post #496

Earlier quoted context omitted.

Because the ability to influence others is more important than 1 or 2 standard deviation in more intelligence. However if we are talking about 10x more intelligence, then that'd be a whole different ball game. Check Alpha Go -> Alpha Go Zero. Gaining orders of magnitude in capability only with slight change to the model, along with a lot of more resource (we can easily throw 100x more compute and data to these models…

>However if we are talking about 10x more intelligence, then that'd be a whole different ball game. Research that the US military has conducted has shown that the ability to influence others declines if the intelligence gap between leaders and subjects is too large. Stephen Hawking wouldn't have been a very good influencer of 80 IQ guys compared to Andrew Tate. 1 standard deviation in intelligence is actually just ab…

I really don't think we can say that research will apply to future AI, given that it was about humans. If intelligent AI exists in the future, it will probably not think exactly like humans. I think you're being overly dismissive.

Re: Pause Giant AI Experiments: An Open Letter

#679

Post by thatfrenchguy » Thu, Mar 30, 2023, 1:32 AM UTC

petilon wrote: post #68

> Should we automate away all the jobs, including the fulfilling ones? You don't necessarily need jobs because you don't need money. What you need is wealth. I am going to leave it to PG to explain the difference [1]: Wealth is not money. Wealth is stuff we want: food, clothes, houses, cars, gadgets, travel to interesting places, and so on. You can have wealth without having money. If you had a magic machine that cou…

Have you ever set foot in a farm, a clothing manufactory, or in a construction site? Maybe mediocre programmers and program managers will be obsolete but none of those three jobs are getting more automated than they already are.

Re: Pause Giant AI Experiments: An Open Letter

#680

Post by evrydayhustling » Thu, Mar 30, 2023, 1:33 AM UTC

reso wrote: post #380

To those saying "this is impossible in our society", there is a long history of scientific fields mutually agreeing to pause certain work until safety and regulations could be put in place. The Asilomar Conference is an example. [1] The idea that "you can't put the genie back in the bottle" is very popular in tech culture, but many have suggested that simply extending elements of copyright or tort law to AI would be…

What about the unintended consequences of pausing research? I know many people working in forestry, cancer, inequality, and the arts producing amazing new work in their fields via AI. What's the cost of delaying advacnces, or a new masterpiece?

And what is the committee that should determine we are ready to proceed? Do we think our representative government is prepared for that task? (The constitution they swear to does not.). Is a conference of academics sufficient? I don't think there is one that represents the majority of world expertise, and certainly not world use cases. Perhaps the courts, but they are designed to seek consistency after the fact, not to extrapolate.

A pause is intended to let alignment research and policy catch up. But can it catch up without the context of developing use cases? When research resumes, can alignment and policy keep the lead, or will we stop and go as adoption overtakes the scenarios we anticipated?

I think a more serious direction for energy and concern is advancing the real safety research and policy decisions evolving around AI today.