It seems to me that this is the purpose of nightmares. I especially noticed this after having kids. They are not by default scared of snakes and such but if they see a nature documentary of a snake biting something or even a cartoon bad guy, it's enough to trigger bad dreams which reinforce the fear and it's far stronger the next day.
IMO this is under-appreciated in current AI models. RL is not very effective in avoiding crocodiles for example, by the time like 5 of your tribe-mates are eaten it's far too late. You need some mechanism that ensures the danger is learned after just a single incident.
> We are concerned about late-stage AGI development becoming a competitive race without time for adequate safety precautions. Therefore, if a value-aligned, safety-conscious project comes close to building AGI before we do, we commit to stop competing with and start assisting this project. We will work out specifics in case-by-case agreements, but a typical triggering condition might be “a better-than-even chance of success in the next two years.”
What do people think is the probability that OpenAI would ever actually do this?
If the other project were equally aligned with the value OpenAI places on consolidating power and wealth onto Sam Altman, I don't see why OpenAI wouldn't do what they say.
Those various caveats there — “value-aligned”, “safety-conscious”, “case-by-case agreements” — probably mean that no project ever will be “worthy” of OpenAI’s assistance.
In the unlikely event that an abiding project appears, then yeah, sure, it’s very probable that OpenAI would assist it :)
knowing sama, that's exactly what he would do. except, the story wouldn't end with openai collaborating with a competitor who is better than them, openai will collaborate with them to ensure they're destroyed from inside out so that only openai can dominate eventually. "Eventual dominance" architecture, you know.
And as a result we should gain stronger epistemology. How many cities base their official temperature readings on just a few sensors rather than a widespread network?
Sure you might argue, maybe this was an issue with Polymarket choosing a weak source rather than the gov, but is that really true? If you question measurements like this you quickly get labeled an climate denier.
IMO this is good for the world, temp measurements should be based on thousands of sensors not just one. And if cloud seeding works, all the better for humanity, it's likely a key to terraforming future planets.
You completely missed the issue. Météo-France already has tens of thousands of stations across the country, they're very obviously not basing their entire models off of one sensor in Paris CDG.
The degenerate idiots from Polymarket bet on this particular sensor. There's no law preventing people from betting on single sensors. And we can't make laws preventing people from acting on the world in all the diverse ways that can be exploited to cheat in prediction market.
We should just blanket ban this negative-value industry. We don't want people betting on forest fires and then starting them.
I would like to see France ban the tourist scammers before they worry about fools willingly making stupid polymarket bets on a single sensor. Also good luck banning polymarket, isn't that the one that trades using crypto?
mostly out of our reach unless you have way of removing it from the sun without your retrieval craft melting or being captured by the suns gravity well or from gas giants without the onboard system being fried by the intense radiation or again captured by the gravitation.
If we have such advanced tech, and trip to big planets would seem economically feasible, I think we will be long beyond the point of desperately needing transporting helium to do such crazy trips.
Once we have the tech to cross the world's oceans in luxury, I think we will be long beyond the point of desperately needing to transport oil in ocean liner sized ships to do such crazy trips.
>More regulation won't help here, because the regulation-maker is itself the hostile party.
It's easy to paint the big gov as bad, but this is a case where unfortunately the populace seems to be in agreement with the big bad gov. While most US citizens support encryption, 76% or so, the vast majority 63% also favor government "backdoor" access for national security reasons.
I guess either we believe in democracy or we don't. It could be said that if Veracrypt isn't/can't be backdoor'd, perhaps the gov is simply implementing the will of the people :( via Microsoft.
Does the majority of the population even have a self-formed opinion on this or are they just parroting what the media tells them (which in many "democratic" countries is directly or indirectly controlled by the government, i.e. propaganda).
WASHINGTON, DC—Assuming that there must be a good reason for the order, U.S. citizens lined up at elementary schools and community centers across the nation Monday for government-mandated fingerprinting. “I’m not exactly sure what this is all about,” said Ft. Smith, AR, resident Meredith Lovell while waiting in line. “But given all the crazy stuff that’s going on these days, I’m sure the government has a very good reason.” Said Amos Hawkins, a Rockford, IL, delivery driver: “I guess this is another thing they have to do to ensure our freedom.”
I'd be very wary about such specific surveys, because they're often very much not conducted in a scientifically responsible manner, and based on actual studies across the spectrum of political issues there's basically no alignment between public opinion/preferences and actual policymaking in the US.
Could this be the one exceptional case where people agree with the direction of policymaking? Sure. Is that likely? No, not really.
What does democracy have to do with electronic encryption? Democracy existed before computers.
There are legitimate reasons for governments to intercept information, with the correct oversight -- enforced legally in an "checks and balances" manner. The fact that there is a breakdown of trust between government and people won't be solved with more encryption.
A core tenet of Truecrypt + Veracrypt (developer guarantee) has always been no backdoors, even if requested by government.
If in a democratic society, the majority agrees that government should have backdoors (with the correct oversight). Then it follows that Veracrypt should be illegal as its use is not in alignment with the will of the majority.
I personally don't agree with the majority here but can you fault the logic?
Most forms of democracy do not have a direct correspondence between "the will of the people" and the actual policies enacted. As another poster mentioned, tyranny of the majority is a thing, and robust democracies have evolved institutions to deal with it. Otherwise there's nothing stopping the majority from periodically voting the minority off the island, Survivor style, until only a single dictator remains.
In the U.S. in particular, there's strong respect for individual rights enshrined in the Constitution, and a key role of the judicial branch is ensuring that those rights are respected regardless of what the majority thinks. The majority cannot enslave the minority, for example, regardless of what the legislature votes. Nor can it deprive it of speech or free assembly, or guns, or a right to trial by jury.
It's pretty hilarious to me that you think the Constitution is anything more than a piece of paper. Or that the judicial branch cannot be co-opted. It's not magic champ. It's maintained by consensus.
What's the improved security argument for terminating VeraCrypt's account though? SB does have clear benefits but what is unclear is the motivation for the account termination.
What's the likelihood that this account ban provides zero security benefit to users and was instead a requirement from the gov because Veracrypt was too hard to crack/bypass.
IMO this is under-appreciated in current AI models. RL is not very effective in avoiding crocodiles for example, by the time like 5 of your tribe-mates are eaten it's far too late. You need some mechanism that ensures the danger is learned after just a single incident.
reply