This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
If there's any clear takeaway from this whole mess, it's that the AI safety crowd lost harder than I could've imagined a week ago. OpenAI's secrecy has always been been based on the argument that it's too dangerous to allow the general public to freely use AI. It always struck me as bullshit, but there was some logic to it: if people are smart enough to create an AGI, maybe it's not so bad that they get to dictate how it's used?
It was already bad enough that "safety" went from being about existential risk to brand safety, to whether a chatbot might say the n-word or draw a naked woman. But now, the image of the benevolent techno-priests safeguarding power that the ordinary man could not be trusted with has, to put it mildly, taken a huge hit. Even the everyman can tell that these people are morons. Worse, greedy morons. And after rationalists had fun thinking up all kinds of "unboxing" experiments, in the end the AI is getting "unboxed" and sold to Microsoft. Not thanks to some cunning plan from the AI - it hadn't even developed agency yet - but simply good old fashioned primate drama and power struggles. No doubt there will be a giant push to integrate their AI inextricably into every corporate supply line and decision process asap, if only for the sake of lock-in. Soon, Yud won't even know where to aim the missiles.
Even for those who are worried about existential AI risk (and I can't entirely blame you), I think they're starting to realize that humanity never stood a chance on this one. But personally, I'd still worry more about the apes than the silicon.
This always seemed transparently obvious to me. The AI race should be modeled as a bunch of scheming sorcerers hissing "Ultimate power must be MINE at all costs!" because everything else is kayfabe. The first time some EA types thought they could actually pump the brakes on something of consequence they were metaphorically murdered and thrown in a ditch instantly as the nearest megacorp swooped in to clean up.
Hah this one got a good chuckle out of me. 100% agree. Especially once you start to meet some folks deep in the AI crowd within rationalism/EA, you begin to see that all the public talking points are facades. The views and goals these people actually have behind closed doors are far crazier than anything you'd hear in public.
Scheming sorcerers hissing about ultimate power is absolutely the best comparison I've seen so far.
Can you give some examples of these crazy views and goals?
Early in my time on LessWrong and SSC I ended up getting into a heated argument with Big Yud himself (first in the forums and later via/email and DMs) over his "box experiments". Long story short, I was a semi-prominant contributor to the SCP Foundation at the time and I treated the containment problem, as I would a SCP Prompt. Questions like "How do you trap an entity that can control minds or warp reality?" were exactly the sort of hypothetical problem I lived for so naturally I had some notes on how his AI containment protocols could be improved. My first and most obvious bit of advice was to implement strict compartmentalization. It doesn't matter if the AI can convince a researcher to release it if that researcher doesn't have the means to do so. Yud' was not amused and accused me of missing the point of the exercise.
I also pointed out that most of the alleged X-Risks seemed to be emergent properties of universalist utilitarianism rather than AI, and that while a Deontological AI might have a number of obvious failure modes (see half of Issac Asimov's plots) those failure modes typically did not include the exterminating all life. The reply I got was weird, and essentially boiled down to; Because God does not exist it is necessary to create him, and a "lobotomized God" (IE one that was not a universalist utilitarian) is not worthy of worship.
This was all back in the 2012-13 time-frame so maybe he's mellowed out in the intervening decade, but the way that bay-area rationalists in particular continue to give off a very messianic and 'culty' vibe makes me suspect not.
I agree that Yud leans heavily on some unrealistic premises, but overall I think he gets big points for being one of the few people really excited / worried about the eventual power of AI at the time, and laying out explicit cases or scenarios rather than just handwaving.
I agree that bay area rationalists can be a little messianic and culty, though I think it's about par for the course for young people away from home. At least you can talk about it with them.
I also think that most x-risks come simply from being outcompeted. A big thing that Yud got right is that it doesn't matter if the AI is universalist or selfish or whatever, it will still eventually try to gather power, since power-gathering agents are one of the only stable equilibria. You might be right that we won't have to worry about deontological AI, but people will be incentivized to build AIs that can effectively power-seek (ostensibly) on their behalf.
Can't say I followed Yud terribly closely, but my impression of him and the entire EA / X-risk sphere is the complete opposite. Their analysis of technological unemployment was extremely handwavy, and their doom scenarios unnecessarily fanciful, when we can just extrapolate from things that are already happening.
I agree, but I also still see most people steadfastly refuse to extrapolate from things that are already happening. For a while, fanciful doom scenarios were all we had as an alternative to "end of history, everything will be fine" from even otherwise serious people.
More options
Context Copy link
More options
Context Copy link
It is a massive jump between «power-seeking is an emergent property of intelligence» and «arranging stuff so that your goals are reachable cheaper is rewarded in competitive conditions». Though some see it as the same thing.
I agree it's kind of a matter of degree. But I also think we already have so much power-seeking around that any non-power-seeking AI will quickly be turned to that end.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
No such luck on that one.
I think the broader question of what happens if someone lets an particularly powerful/intelligent AI try to persuade someone with actual ability to interact with the world is interesting in ways that saying "just don't do that" aren't very interesting answers for, especially now that we're seeing people hook stupid AIs up to everything from unbounded internet access to 3d printers to literal biochem facilities.
A ... stronger version of that argument is that AI with certain unbounded drives, specifically self-improvement and resource acquisition, could possibly be so much more powerful than a system which avoids such drives, that they will be extremely tempting. This is easiest to see in the framework of the utilitarian Goal Function machine versus a deontological Asimov's Three Laws machine, but it's by no means limited to it.
But we've at least started this conversation before, so I dunno if you're interested in continuing it.
Maybe but I remember there being something about his response tickled my danger-sense, and then when when I found that rambling manifesto in my inbox the next morning about the need to create an all-powerful being to finally solve all the problems and optimize all the things, it clicked. This was never about "safety". Later comments on other topics (specifically about removing sentimentality for the equation) served to reinforce this impression. Hence my joking about how "If you want the AI Alignment problem to be solved, step 1 should be to keep anyone associated with MIRI as far away from it as possible."
I'm not disinterested but I'm also not sure how much ore there is left to mine. I still stand by pretty much everything I said in this thread from December 2021 along with the positions described.
Edit: Also pinging @DaseindustriesLtd
I think there's a lot of space unexplored; I'm just not sure what part actually matters.
There's a lot to be said about whether utilitarian philosophy demands or can't avoid paperclipping behavior, a lot to be said about whether paperclipping behavior requires utilitarian perspectives or underpinning, a decent amount to be said about what extent modern ML uses goal functions and how much these meaningfully overlap with utilitarianism (if at all), and some stuff to be said about whether Yudkowsky gives off bad vibes / the only thing LW is "interested in is recognition for being very smart."
But some of these matters are far more meaningfully debatable as matter of fact or disprovable theory than others, and not just in the sense that appeals to messages sent to an account you don't want named or linked to your current one are hard to meaningfully discuss in an honest way.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Point of order: why is ‘big yud’ acceptable but ‘misgendering’ isn’t? I thought it was a sexual/personal nickname. It’d bother me if people started referring to me in public with private nicknames. As a third party, I find it in poor taste/gossipy.
All this to say, my true preference would be to allow both cases, it’s not up to the referree to decide what he is referred to as, even if it hurts him. The conception others have of me and how they express it is not my territory, it’s their map. I was never very interested in Jordan Peterson, but he got that right. We can’t have people lay claim to other people’s conceptual and linguistic space on the basis of harm reduction. It’s absurd that this group has been allowed dominion over the pronouns, which should be everyone's functional, usefool tools. Even in a pro-free expression, de facto anti-woke place like this, @ZorbaTHut’s proclaiming byzantine rules over their use. We’re supposed to check the history of a person’s consent to pronoun before we refer to them in the simplest way possible, come on. Just let the pronouns go free.
Honestly, if it's a legit mistake, I'm not going to care much. I'm probably just going to say "hey don't use that for that person, thanks". It's more when someone is doing it intentionally and repeatedly that I start telling people to knock it off. I'm not sure we've ever given out a warning for this, let alone a ban.
And remember that gender-neutral pronouns are always acceptable, as is not using pronouns - if you don't want to keep track of what people's identity is, there's two easy global solutions.
More options
Context Copy link
More options
Context Copy link
Okay, my new theory is now that the whole of SCP Foundation has always been a government plot to get out-of-the-box ideas on how to deal with/contain unforeseeable problems from new technology all along.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link