Sorry, I was a bit confused by:
If Google runs Kimi on its own hardware, Google is responsible for any actions it takes. If a private individual runs Deep Seek locally they are responsible for any actions it takes.
In both these cases, the person deploying is also the owner of the hardware.
I feel like there are mainly two potential issues with this:
if one thinks AI has huge potential for good, then this would 100% significantly hamper that, because it creates the classic extremely risk averse “cover-your-ass” sort of incentives that have similar effects on many other fields already. This is really a divide about how pessimistic one is about AI outcomes, and thus how much utility is lost by limiting them this way.
I don’t like the “company is at fault for things run on their hardware”. If a user rents hardware from Google to run a model for their own purposes the liability should be with the user, not with Google. Otherwise we incentivise and require levels of surveillance of the companies on their own users that I think instead we should discourage.
This is not centrally the case. You don’t need to go for the median to draw a sensible boundary and then retreat from it to the inside, with some margin. There will remain a lot of variation, on both content and quality, but you won’t risk the failure modes of non-centrality.
I mean, I really have no objection to this—any kind of understanding of a whole distribution is always better than just understanding of individual simplifications of it (be average, median, or what have you). And it’s not like most of this stuff is literally something you can draw on a single 1D numerical axis, except for some very rare and specific policy positions such as “what tax rate in % should be applied to people above £100k” or such.
I was more gesturing at this “try to grasp what the common position is” concept. If you have a huge bimodal distribution where the movement is split almost 50⁄50 between two major interpretations of the belief then sure, both of those are more relevant than the rather meaningless mid-point. Apply pinch of salt and common sense as necessary.
The criteria that grade arguments as sane/thoughtful/well-articulated should be suspect, because something might be such from your point of view, but not from inside the target worldview.
Ultimately yes but there’s only so much relativism I can engage in. If someone posits a completely different worldview from mine down to “empirical knowledge isn’t valid”, what can I even do? Maybe it makes sense for them, but for me it’s insanity and as such disqualifying. This is different from “ok they share with me the basic principles of how the world works and how knowledge is accrued, and merely have different information, so if we engaged in enough mutual exchange we might come to an agreement or at least understanding”. Otherwise you’re not steelmanning, you’re just passively acknowledging any idea as just as good as any other, because for any idea there could be someone who thinks it’s perfectly coherent and sensible (for starters, usually, the person who voices it).
The purpose of steelmanning, in the (more useful, I think) distortionary variant of its meaning, is to build your own ideas regardless of whether they accurately represent and should be attributed to their original source/inspiration.
Steelmanning is used a lot of time as a way to engage with ideas/groups by giving maximum grace. Basically trying to play devil’s advocate to make sure that your own criticisms aren’t facile or purely the result of blind outgroup distrust, but grounded in genuine rational thinking. It’s not how you recommend to use it, but people do this, so I’m responding to that mainly.
I thought we were talking about a shockwave, so ultimately a pressure front? That’s the part that travels several times around the world. I mean obviously with plenty of volumetric loss/dispersion on the road.
I guess I’m thinking about whether it could be an amplifier of some kind—your immense energy beam hits something compressing the matter to that kind of critical extent and thus adding its energy to the process. But I suppose ultimately it’d pretty much cost as much to compress as is released on evaporation so probably not a particularly useful addition.
Shouldn’t the scaling be yield^{1/2} if you’re considering energy radial flux? But also, on these scales we should consider whether the atmosphere “channels” the energy as it bounces between the ground and the upper terminal points thus achieving something more like a ~yield scaling?
I had never thought that strangelet growth dynamics would be dominated by them being exothermic, so that they heat protons away from them and that hampers their growth.
About tiny black holes, another possible use for them would be simply as bombs themselves, right? Evaporation should be catastrophic and highly energetic for unstable black holes, and basically free up
I’ve been pretty clear though that the fleshman isn’t about inventing ideas, it’s in fact meant to be more empirical than the steelman and something you have to go and observe the movement to extrapolate. The point I’m trying to convey is more about the “median” aspect. Some people are quite content to go on X/Twitter, find the Dumbest And Craziest Member of the outgroup, and point at it as a sign that the outgroup is Obviously Bad. This is technically “engaging with the outgroup”—the craziest member is still a real member! But it’s also a form of cherry picking evidence. Similar things can happen with the steelman—out of epistemic humility you might go and find the most sane, thoughtful representative of the idea you’re arguing against, someone so well articulated that you have a hard time finding fault in their version of the idea at all. That’s also engaging. But should you let that kind of thing discourage you from ever opposing that group again, on account of, they do have a point after all?
This is the thing I’m stressing. The opinion of the typical member is something you can only glean by observation, not pure thought, and is the one thing that’s going to matter the most. A classic example of this would be seen in many religious beliefs—the most educated theologians and scholars are not a good proxy of how it feels like to live in a society dominated by those beliefs. The typical believer is.
I don’t think that people who are ok with/optimist about AI development all count as e/acc. e/acc for example has an official manifesto, and it includes sentences like:
e/acc is about having faith in the dynamical adaptation process and aiming to accelerate the advent of its asymptotic limit; often reffered to as the technocapital singularity
Effective accelerationism aims to follow the “will of the universe”: leaning into the thermodynamic bias towards futures with greater and smarter civilizations that are more effective at finding/extracting free energy from the universe and converting it to utility at grander and grander scales
This isn’t just “we think AI will be great and bring more prosperity and all those worries are overblown”. It’s explicitly about making the Singularity happen ASAP. AI optimists like, say, Yann le Cun don’t seem to believe in a Singularity—they think AI will be just another tech.
I mean, most of them sound great if you’re an edgy 15 year old on a pseudo-philosophical kick.
I don’t think this is quite what e/acc is motivated by—many of them are more openly focused on ideas of AI as successor/evolution of humanity. Well, Beff at least seems to be. I’ve definitely seen the position you describe too but I don’t remember if the people espousing it self-identified as e/acc.
I feel like e/acc is ultimately a pretty small subculture of mostly Extremely Online people. Beff Jezos is as you pointed out essentially a grifter—his whole company seems based on flimsy theories and his ideas on entropy are incredibly poor and flawed. As it often happens, these people are visible because they can be loud and because their ideas are so cartoonishly villainous they draw attention. “If the AI kills us it means it’s better than us so it’s fine, evolution must take its course” is a line the average superhero movie scriptwriter would hesitate to pen for fear of being too cheesy.
I don’t think they’re too consequential. None of the really big shots at the main labs are e/acc that I know of. But I do think there’s value in criticising them anyway since their ideas are so bad. Just in moderation and only when they already pop up, lest one boosts them by Streisand effect.
The proponents of anything cover generally a whole spectrum. The point is more “pick something that is meaningfully representative of that spectrum”.
I mean my point is that having a stronger-than-justified belief in the healthiness of apples leads to certain conclusions that are wrong because the belief itself doesn’t correct map to reality. And knowing the ways in which that happens is important to predict what mistakes they might make.
I suppose it sounds a bit gross! I think the “they’re made out of flesh!” old sci-fi story was in my head when I thought of that.
My point isn’t that the steelman has no purpose. They’re tools for different situations.
As you say, “Is Marx still to be taken seriously?” requires a steelman. If Marx passed muster then that might mean we’re supposed to build a new rationalist socialism that is as close as possible to that idealised and strengthened version of the argument.
But “should I support current Marxists as they exist” requires a fleshman.
and where it is, I don’t think the “steelmanning” as concept tends to sneak in.
Personally I’ve seen this happen often. Though often with the other end of the political spectrum. People engaging with the smart version of an idea even when the median supporter is very, very dumb and there’s no way anything that will be done in their name even comes close to that smart version anyway.
That’s slightly different as I feel like the core of that is “do you have a good theory of mind of your opponent”. If I had to pretend I’m a believer of something I’m not I wouldn’t necessarily go for the median—I would go for what I, specifically, would be like, if I bought into the core premise. It’s the perspective I can empathise with the most, the minimal deviation from my current self.
I mean, yes, houses are mostly as valuable as the location they are in. That’s true but it doesn’t really deny the problem. Rural houses aren’t very valuable simply because there’s just not a lot of economic activity going on in rural areas in industrialised countries any more.
Sorry, my bad probably, I’m not very well-rested so I may have not registered that properly.