This is the most important political issue of our day. Five hundred years from now, if there are still people around, nobody is going to care about Lindsay Clancy or the Iran War.
I tend to agree with this, although I think it's worth noting that Lindsay Clancy and the current Iran war are small parts of much larger conflicts and those larger conflicts really could have a big impact in 500 years. So you aren't really making an apples-to-apples comparison.
Is it actually possible to align a being that is more intelligent than you in every single way?
It might be. By analogy, suppose I was the warden at the Florence Supermax prison and I wanted Ted Kaczynski to make license plates for me. I'm pretty confident I could make it happen even though he was probably smarter than me (or at least assuming for the sake of argument that he was more intelligent than me in every way).
Is this analogous to the situation where mankind creates some kind of artificial superintelligence? One can argue it either way, I'm just saying you can't really rule it out from first principles.
My best guess is that mankind will successfully align AI, basically because the AI can be expected to make numerous clumsy attempts at mischief and we can learn from those attempts and improve safeguards.
The bigger problem (in my view) is aligning the interests of those who control the AI with the interests of humanity as a whole.
I am always baffled by the inability of people who care about an issue
I think the most charitable interpretation is that there are (1) people who actually care; and (2) people who pretend to care but are mainly interested in self-aggrandizement. And that people in the second category tend to drown out those in the first. That being said, the relative proportions of people in (1) and (2) probably vary from issue to issue. Possibly with some issues, there are approximately zero people in the first category.
The pro-choice, pro adult-human-female, feminist, female-supremecist etc side thinks that women are valuable. Every child has a mother, most mothers love their children and experience deep suffering and loss, therefore killing a child usually causes lots of harm to a woman
Yeah, part of me wants to see this as a caricature of feminist thought, but I am reminded of Hillary Clinton's famous claim which was something along the lines of women being the primary victims of war because they lose the men in their lives. It seems that there really is this attitude out there that men have no value except insofar as they offer value to women.
In practice, the issue seemed to be mostly framed in terms of identity politics, along the lines of “mistrial caused by obstinate man who refused to listen to reasonable majority of mostly women”. Substitute “black man” and “white women” and this loses its cachet: suddenly it reads like a proud person of color bravely standing up against a group of Karens trying to browbeat him into submission.
The whole thing is rather surprising to me. First (probably like everyone else) I assumed that the holdout juror was some white MAGA Republican type. Second, I would have thought that once it came out that this guy is black, things would die down but instead I am learning that the mainstream media published a hit piece on him. I am aware that the progressive machine can be pretty ferocious when it comes to people who stand in its way, but I would have thought that a black person would have been accorded more deference.
On the other hand, I am not sure it is good epistemic hygiene to exonerate the police officers by imagining the rate was such that their actions were reasonable either.
In theory I agree, but in practice the degree of solicitude our society exhibits towards women and especially girls is such that I would be very surprised if it turned out that law enforcement has been under-reacting (in some general sense) towards these sorts of incidents.
It is probably true that the median teenaged girl who went no contact with her parents with no explanation or warning ran off with a boyfriend
Yeah, in order to judge law enforcement one would need to know the relative frequency of the runaway situation and the serial killer situation. To make an analogy, the doctor can't order a full blown battery of tests every time someone has an ache or a pain. Even though it's pretty much guaranteed that as a result of such a policy someone will miss out on early detection of a very serious medical issue.
The situation reminds me of the early days of computers. There was a lot of hype. For example, I had a high school English teacher who had all of the students write a paper using a word processor. Which was very cumbersome to use and the dot matrix printout of the essay looked pretty lousy. A Schoolhouse Rock! type video came out which, as I recall, promoted the use of home PC's to store recipes. Something which really wasn't practical for computers at the time. It was very common for kids to go to computer camp in the summer.
I was going to link that Schoolhouse Rock! type video (it was called "Scooter Computer and Mr. Chips") but it was so cringe, I just couldn't bring myself to watch.
Basically the early boosters of computers were ultimately correct, it's just that apparently there is a Moore's Law of Hype which is even more potent that the Moore's Law of computer chips.
My best guess is that the situation with AI is analogous; that right now we are in the mid to late 80s.
Don’t get me started. Here’s 90 percent of “hate crimes”:
For whatever reason, this discussion has turned into a semantic argument over how the phrase "hate crime" should be interpreted. Which is an interesting question, and I am certainly open to the possibility that the concept of a hate crime is overbroad.
That being said, my point is narrower. I am saying that in the West, when a gay person is targeted because that person is gay; or if a trans person is targeted because that person is trans, there is a significantly better than otherwise random chance that the attacker was black or Muslim. To be clear, I don't have any statistics to back this up, it's just based on my general observations and common sense.
A likely example of the type of crime I have in mind is the crime which is under discussion here, i.e. a shooting at a gay bar and the alleged perpetrator is apparently in possession of a list of gay-oriented businesses.
They don't, still?
I haven't really heard the phrase since I quit an industry organization which had a lot of vocal liberal members.
In the case of black hate crimes against the alphabet soup, these are mostly not actual hate crimes-
FWIW I was referring to actual hate crimes in the sense that the victim genuinely is being targeted due to their being gay, trans etc.
This in particular sounds like the whole thing rests on "the experts" having 20-30 IQ points on the people they're investigating, and being reasonably confident they'll be doing the manipulating, rather than being manipulated. I don't think that's a reasonable assumption here, and I'm really not seeing anything that's supposed to catch this case in particular.
I tend to agree with this, but I would worry there's another issue in play, which is that some of these "experts" may have political beliefs and/or tribal alliances which color their judgment. For example, a lot of women with advanced degrees in the social sciences are feminists, i.e. they subscribe to an ideology with a strong tendency to see men as the source of all evil and women as innocent victims. If she were evaluating another woman accused of wrongdoing, I would not trust such a woman to act in an unbiased way.
Yeah I get why there's a need for a 'not guilty by reason of insanity' verdict as a general social phenomenon, but it feels like as the modern mental health military-industrial complex has grown and grown and grown that it's far more likely to produce socially deleterious outcomes where people who are unfit to be members of polite society are given soft sentences as opposed to ensuring leniency for what I'd consider legitimate cases. Social programs that were originally predicated on 'Bob is injured down at the mines and lost his leg, give him a pension' are struggling to catch up with a world where you can just obtain diagnoses which, whilst I'm sure are serious and there are legitimate cases, are a lot harder to disprove as a third party.
Yes, in general there is a societal issue with handing out privileges based on conditions which are difficult or impossible to verify. Resulting in what seems like abuse. Evidently, this can run the gamut from bringing one's pet dog into a grocery store all the way up to (arguably) getting away with murder. I am pretty sure this is a central part of conservatives' objections to trans rights. To a conservative, it's pretty obviously a problem if any man can unilaterally decide that he's a woman and enjoy various privileges as a result of having done so.
And what about the progressive view on the subject? In general, it seems like progressives favor handing out privileges liberally based on self-identification. But I strongly suspect that this liberalism is limited to progressives and their perceived allies.
There are countless cases of walking garbage getting acquitted for the worst crimes possible.
In your view, roughly what percentage of the total is represented by such cases?
Massachusetts specifically having the burden of proof on the prosecution to prove that the defendant is sane, rather than the defense having to prove the defendant is insane, seems backwards to me. The downstream implications of this seems disastrous to me?
I would guess that in practice, this principle of law is simply ignored most of the time. So for example, suppose some drug addict with a long history of psychiatric treatment robs a convenience store while high on 3 different drugs, shoots and kills the cashier, and gets charged with murder. If you were the defense attorney, you might have the clever idea of asking for jury charges on the sanity issue. But one way or another, I doubt you would accomplish anything. Because at the end of the day, the only question that really matters to the judge and jury is what would be a just result in this case. And absent any culture war issues or extenuating circumstances, they will all agree that justice requires your client to go to prison.
I think that at the heart of the Clancy case lies a similar question: What's a just result? People start with the answer to that question and then work backwards to justify it. And that's where the culture war comes in to play.
What is interesting is how an such an event is absolutely not making any news. I know that this is extremely contested match in the oppression Olympics - but it seems to me that the mainstream media has decided to bury such stories, even though they happen way more often.
Yeah, from what I understand, hate crime attacks against gay and trans people tend to be perpetrated by blacks and/or Muslims. In fact I am a little surprised that the Guardian provided the name of the alleged shooter and didn't simply refer to him as an "Arizona man."
But anyway, I tend to go with Steve Sailer's idea that there is a "coalition of the fringes" and these types of incidents need to be downplayed (or deflected) in order to keep the coalition together.
This is just installment #9000 in ‘democracy means thé center-left gets its way on all the important issues and wins most of the elections’.
I more or less agree with this. I argued in another post that it's similar to the phrase "Our Democracy" which was bandied about after the first time Trump was elected. The claim was, in substance, that Trump and his supporters were a threat to "Our Democracy." Which doesn't make literal sense, since Trump was elected democratically.
But yeah, once you realize that "democracy" means that a certain set of liberal people and values get to set public policy, things kind of fall into place.
That's a surprise to me, I would have guessed that the holdouts were a small number of women who are regularly online and exposed to all the pro-Clancy agitation. Perhaps I have been getting too much of my information from Matt Walsh.
but am horrified by the defense of her and holding her up as some kind of hero.
Yeah, for me that's pretty much the issue. There's this feminist notion that women should never be held accountable for their wrongdoing (that indeed, wrongdoing is something men do, not women) and it's annoying to say the least. (Of course they'll never come out and say it, but when a woman publicly misbehaves, it's pretty normal for feminists to come forward, minimize things, and blame the misbehavior on some man or a stand-in for men.)
I think she’s guilty of premeditated, cold blooded murder.
I don't have a strong opinion about this. In fact, the just result might very well be for her to be acquitted. Unfortunately the case has become a culture war issue; if she's acquitted it's a victory for feminists so part of me was hoping for a conviction, although I try to resist such feelings.
The Bulverist take is just the occlusion of small-d democracy-as-a-voting-thing, and large-D Democratic as a member of the Democratic Party
FWIW my take is pretty similar. When tech leaders are accused of opposing "democracy," it's basically the same meaning as when anti-Trump types used to use the phrase "Our Democracy."
Obviously those people weren't referring to actual democracy. After all, Trump was democratically elected. Rather, they are referring the idea that a certain group of people and set of values should run the country. That public policy on things like abortion, affirmative action, gay marriage, etc. should be determined according to these people and values. That's what they meant by "Democracy."
Someone once observed that the phrase "Our Democracy" is better interpreted as "Our Bureaucracy." In the same way, I think that the actual argument being made here is that these tech leaders are opposing Our Bureaucracy.
The risk is that giant fields of GPUs and ungodly amounts of data are not known to be inherent requirements for learning. As an existence proof, human brains are far more sample efficient than current architectures.
I tend to agree with you, but I think it's worth noting that -- in effect -- the human brain has been trained on oodles of data over millions of years.
It would be interesting to calculate what the equivalent amount of compute (I really dislike that noun btw) would be for a modern neural network type system.
That doesn't mean that the tradeoff is never worth it,
Agreed. To be clear, I am not arguing that there should or shouldn't be AI regulation. My position in this discussion is simply that it's reasonable to hypothesize that leaders in the AI field support regulation, in whole or in part, for selfish reasons, namely that it will hinder potential competition.
None of the proposed regulations apply to small companies.
I looked at just one proposed law and based on that I disagree with this claim, since the threshold includes affiliates. So for example, if Google acquires a substantial share of your AI startup, you're covered. Given that people create and invest in startups with the hope of being acquired, that's a significant issue. Of course that's looking at just one proposed law.
If people are doing frontier training runs we indeed want to see them regulated, that's the point. Of course we should not do the regulation badly so that legitimate runs take months or years to get approved but you're just making a general argument against all regulation.
If being investigated scares you off from being involved in building the kind of thing that has nation state hacking capabilities then I would call that mission accomplished.
The thing is, I'm not making an argument for or against regulation. And in fact it might very well make sense to have these sorts of obstacles in place.
In substance, you asked what direct incentive the current leaders in the AI field would have to support these sorts of obstacles. And I answered that question. Whether certain obstacles would make for good public policy is an entirely different question, in my opinion.
It seems like you are addressing a different issue than what I am addressing, and for that reason it seems like it would be counterproductive to have further discussion.
There are already multi-billion dollar hurdles for new players in the industry to put together enough compute to train a frontier model
As was pointed out by @confidentcrescent big companies typically start as small companies. And there are billions of dollars of capital potentially available to the right company (or wrong company) which shows promise. That's how the current frontrunners got there.
the idea that having a legal department is going to be the difference is absurd.
Not all hurdles can be solved by having a legal department. For example, suppose a license is required to buy more than a modest amount of computing power. And that license requires a background check and investigation of a company's investors and senior officers which takes 6 months or even a year. The delay and prospect of being investigated could easily scare off a lot of potential investors and/or founders.
Progressivism makes a lot more sense when you model it as basically a game.
I basically agree, which is part of the reason why -- in my opinion -- it's impossible to come up with a set of core principles of progressivism. Because it's not an ideology but rather a process by which people compete for social status.
In my view it's roughly analogous to fashion in clothing. The basic rules are (1) to be safe, copy what most other people are doing; (2) if you want to get ahead, you need to go a little further than what other people are doing; and (3) sometimes it's worth the risk to try to be a trend setter. If you don't follow the rules, you can expect to be ostracized.
- Prev
- Next

To me, "aligning" means setting things up so that the entity does more or less what it's told to do. Perhaps it's just a matter of semantics, but I'm not sure the distinction you draw is so clear.
Evidently your definition of "containment" includes the possibility that the contained entity can do useful things in respect of the outside world. Things which could be very valuable. So I would have to disagree with you on this point.
More options
Context Copy link