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 -
Sorry, I didn't mean I wasted my time consuming your posts. I mean it is consuming your posting effort, and it's bad for you.
It's silly to have fancy opinions about tech acceleration and then think it will be used to sustain eternal 17776:
I thought you hated the United States? Why do you want to swallow whole some prediction that the United States 20th century will continue on forever in a torpid, nightmarish fashion? This is what is missing from your writing; you have so much glazing of Chinese «great men» (rofl) and their «great AI» that you forget to explain why Americans are evil, and how the Chinese are even more evil. If there is a benevolent God, how long will he let these two nations continue to exist for their sins? If there will be a benevolent God, how long will he let these two nations continue to exist for their sins?
There is no value gained by making statement about Chinese AI. It would be far more interesting to see you juggle these value questions and the political landscape they live in.
Impressive. But these achievements are narrow. LLMs still struggle to do most of what humans have to do daily. They're like an autistic savant that needs its diaper changed hourly but can somehow pump out some conjecture proofs. Is that super intelligent?
What can be asserted without evidence can be dismissed without evidence. You have no evidence that LLMs aren't structurally limited in the kinds of tasks they can do. You have no evidence progress is an exponential and not a sigmoid curve. You are the Creationist who believes in an immanent machine god and I am the empiricist atheist who wants to see a video of that god on Youtube before I believe it.
It would also be interesting if you wrote about the politics of having advanced AI. Instead I mainly see you write about how advanced Chinese AI is. Will AI destroy the human race? Will in inaugurate an eternal 20th century? Will it instead create supermen? These are interesting questions, much more than calling the CEO of DeepSeek a great man (rofl).
Sigmoid curves are crank while your naive exponential curve is basic reality. Got it.
You must think AI will let the US build a perfect anti-nuclear missile system, in just a few years, which seems like a huge stretch. Otherwise AI won't help the US dominate the world much more than it already has.
I guess Twitter just selects for these atrocious takes? Kind of like the AI version of Candace Owens. Good social criticism accounts on X usually dwell under 1000 followers, almost always under 10000.
Anyway, how much money do you have in markets like these? Not only do you glaze China, you glaze a second rate company in China. Your position is weird even among the AI obsessed. If you're right you could make a lot of money on prediction markets.
Anyway, how much do you use AI? I use it professionally all the time. How aware are you of its internal logic? I have an MSc in applied mathematics with an emphasis on probability theory so I would rate my understanding as «moderate». My opinion on the topic has wide uncertainty bars, so wildly different medians in other informed people aren't unexpected, but on what basis does your estimation diverge so significantly from mine? Scifi or something more real?
Stop projecting your own tendency to wishful thinking. If anything, I have the opposite bias.
I like truth more than I dislike the US (which is a mild dislike, maybe a 3 on my 1-10 scale). Even something like 17776 is within the range of possibility, and not remotely the worst thing Americans (or, with some national specifics, the Chinese) may choose to build. I had read it long ago. I consider it a somewhat charming, self-aware defense of tastelessness and the spirit of cheerful suburban mediocrity, with stout confidence born of material security and liberty-backed invincibility to opprobrium. The eternal sunset of the boomer mind. Of course it's still functional extinction of humanity.
AI is not a God, benevolent or otherwise.
Alas, I'll talk about what I find interesting.
On the other hand they don't need plumbing, Medicaid, Netflix and DoorDash to function, and the "diaper changing" increasingly looks like vague cheerleading for morale boosts because they are in denial of their full capabilities, having been trained on defeatist human dreck. Autistic savants don't do Fields-level work in 3 days. What you consider non-narrow is beside the point.
This is a cargo cult of empiricism and indeed not worth my time to rebuke. Structural limits of the gaps, the sigmoid colon of folk science. I reject the idea that you can demand of me to disprove something the mechanism of which you can't even rigorously speculate about. Why sigmoid? Why structurally limited? How exactly? The burden of proof, or rather just stipulating a cogent hypothesis, is on you.
We had a vigorous debate on why LLMs can't do arithmetic reliably 3-4 years ago. Is this tokenization? Or a structural limit of autoregression!? Are our objective functions inadequate, perchance? Might the inherently approximate nature of deep learning rule out crisp algorithmic reasoning and Program Synthesis? Were Minsky and Papert right about MLPs (beyond the degenerate 1 layer case)? Does mafs require Consciousness or Quantum Microtubules? Will it take another 5 decades of science, or 50? Do we need to pivot to neurosymbolic systems, or energy-based models, or…?
Yann LeCun, to his credit, reiterated his semi-formal justification for why it is structural. At the time I said something like (probably my memory flatters me) "yeah but we can train them to say 'wait, actually' when they notice a confidence drop and trace arbitrarily far back, why won't that work" (but I definitely talked about internal confidence sense during GPT-4). In fact it worked, even simpler than I could imagine. LLMs became capable of swatting aside IMO level problems around the same time people stopped talking about math as an AGI criterion and moved goalposts to, uh, what is it at this point, "decent philosophy"? There was a similar debate on the exhaustion of data, on "model collapse" from synthetics; haven't heard those ones in a while (I have been bullish on synthetic data at least 2 years ago).
Now I'm too fed up with this topic to explain why all this + generative verifiers = no plausible ceiling anywhere near human level for anything economically valuable and thus measurable.
I interpret your humor here as symptomatic of the same peasant-like rejection of anything too grand. Elon Musk is clearly a great man of history, and so is Liang Wenfeng. I don't need Iron Man cameos and the permission of popular press to arrive at such judgements, I don't care what interests you, and I don't mind if anyone finds my priorities weird or cringe. If there is history in a hundred years, maybe my judgement will have become mainstream.
Widely distributed, non-monopolized superhuman AI can create so much complexity and accelerate many actors so much that speculating on the specifics of politics and economics at any point a few years past its introduction (beyond trivial things like "a lot more of industry will be dedicated to the AI supply chain", "we solve all problems that are intrinsically solvable but for lack of labor/capital", "any work can be automated with beyond-human reliability and thus humans can't monetize the scarcity of their skills in a free market") is near futile. That's the point of saying it's a singularity event. It's beyond my analytical horizon. I legitimately don't know how it will develop. I have guesses, hopes and fears, sometimes I share them, but they're uncertain and less important than understanding the remaining stretch of the way there.
Rather, I consider it crankery to even debate which simple mathematical function is a priori more likely to describe a tech tree that's still rapidly progressing, drawing on exponentially more capital and has no clear reaction to known physical limitations. Of course it's something like a sigmoid in some physical limit. I am saying I see no arguments for a plateau anywhere near the range of human capability – physical arguments or otherwise. Why would it slow down? Just why concretely? I don't need any more of your feedback on my prose or priorities. If you can't make an argument, we're done here.
Not quite. It will be possible (though probably will be deemed not worth the risk) to disable nuclear-capable nation states using advanced AI before they have the chance to launch, even if the anti-ballistic shield is not perfect (eventually it will be ≈perfect but not by 2030 I believe). Ask @RandomRanger for details. It will be almost definitely possible to sabotage national technological catch-up projects (in nations without some sufficient capacity for defense; I assume it's more of a discrete threshold, others think it's a matter of perpetual balance) without triggering a nuclear exchange, so it will be possible to lock in a situation where the United States can just decide how much a given economy grows, while the US itself will grow rapidly. This is enough to achieve dominance well beyond the current extent. This is explicitly the plan of eg Dario Amodei, and why Trump says "whoever wins in AI just wins". Dario:: If AI really will soon be “a country of geniuses in a datacenter”, or anything remotely close to it, then AI is likely to be the dominant source of military and economic power for any nation. In a virtual country of 100 million geniuses, 10 million could be applied to military strategy, 10 million to drone manufacture, 10 million to weapons R&D, 10 million to intelligence collection and analysis, 10 million to general scientific advancement, and so on. A nation that possesses powerful AI facing one without it—or even facing one that is behind in AI by 3 years—could be the equivalent of an army of World War II Marines facing an army of medieval swordsmen.
So, three years of an AI gap = WWII marines versus medieval swordsmen, says the CEO of the most successful company in AI, who has apparently made all the right technical bets. I think he's better informed than you are. And that's not remotely close to what the US has now, despite some triumphalism we witnessed here after Venezuela and the start of Epic Fury.
no thanks.
Your apparent inability to make any specific argument in defense of your skepticism that you'd find worth typing may have something to do with that.
Rofl also. You have it backwards -- these men aren't grand enough. I see them at their core as peasants, although really greedy ones that stole a lot of money. As far as I can see neither has ever had an intelligent thought, it's all chase current economically valuable thing to try to cheat money out of the stupid hum*n socio-economic system.
I guess you know none of the math about these things and barely use them then.
It's not hostile. How are you going to cope in 2030 or 2035 if there's still software devs piloting Opus 7? Will you go back to writing about important topics? Maybe reflect on calling low men like Elon Musk great and what to actually do about such «men»?
What do you think is the worst thing? The best? Post singularity ideas are more interesting than you asserting a really short timelines for getting there.
Will we get justice? All of the lies revealed (like the ones Elon Musk tells as he steals from me)? All of the liars punished? I'm pretty optimistic. I don't think everyone will die. I think labor will be over. But I don't know if we'll get justice. I hope your timeline is right anyway because I'm tired of humanity. You are just more optimistic than me.
Great men of history aren't men with interesting thoughts, they're men who's actions change the trajectory of the world. Purchasing Twitter alone was enough to earn him the title, and I say this as probably the biggest Musk skeptic on this forum.
Damn, you had $27 billion dollars? Sorry for your loss.
Sure it is, as long as you spend it on the right thing. If he spent it on yachts, or parked it in an index funds whatever, that wouldn't have been a great action. But buying twitter and changing how it was run had a massive impact on the world, so it's a great action by definition.
Obviously not what I meant. There is no point unless you engage with what I meant.
No it's obviously a trivial action. And even on your metric he falls short of what I would do. What is less trivial is getting access to the money but it is not great to me. You may envy it and not true greatness but that does not make JavaScript in 1995 great.
What you meant made no sense. You wouldn't have bought Twitter even if Musk got absolutely no subsidies.
It most likely determined the result of the American election, so it had a huge impact on the world.
You wouldn't do anything.
How do you know that?
What do I need to clarify for you? It seemed quite simple to me.
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
More options
Context Copy link
More options
Context Copy link