Robin Hanson

@FleekFuturist Might not be the worst scenario

@robinhanson here’s an oddball react for you though: Plan B, civilization cannot be saved, collapse is inevitable, but we organize highly insular, highly fertile, highly capable subgroups, that can outcompete with the traditionalists, rebuild after the collapse, build back better 🏋️

If you are high enough status, it can pay to join alliances and play politics. But most are wisely advised to avoid messing with governance, to try to hold leaders more accountable, as leaders fiercely retaliate against such efforts. overcomingbias.com/p/on-politics-…

@Aella_Girl Blame is a big thing, so avoiding it gets big priority

it's weird that 'innocence' has such a positive valence. Not being innocent - being exposed to the world, knowing more, having experience - feels *good*. Are pro-innocence people those who feel like exposure to the world has been negative for them?

@warty_dog @danboneh Moore's law will only apply to the usual computers with huge demand driving innovation.

@robinhanson @danboneh in near term yes, in long term moore's law should solve this

.@danboneh tells me: quantum computers can't use quadratic search advantage, so main advantage is quantum sims & factoring to break old crypto. Their speed & size is much worse than usual computers, so this will be small specialized device industry.

@warty_dog @danboneh Fixed gains of usual computers means need huge scale to get that quantum advantage, which also needs infeasibly large space scale.

@robinhanson @danboneh why can't use

I live near DC, and when I travel to West coast US, I am temporarily a morning lark, while when I travel to Europe I am temporarily a night owl. Seems to me morning larks get more solitary work time, while night owls have more social fun.

@zdch Only being able to think at one level of analysis is indeed quite a limiting factor.

@robinhanson I agree, but I also think this regularly leads to an inability to operate at the level one “take the statement at face value” level of analysis, which is often extremely explanatorily valuable.

To the contrary, we can learn much by studying ugly hidden motives. elephantinthebrain.com

@schulzb589 But isn't that a vastly smaller industry?

@robinhanson CAD is next. Part design is quickly going the way of coding.

“The pessimistic case starts from the premise that software development is different from the rest of white-collar work. Coding involves huge amounts of training data, a relatively limited range of possible outcomes, and outputs that can be objectively evaluated—all of which makes it ideally suited for AI automation. That isn’t true of all knowledge work.” theatlantic.com/economy/2026/0…

I have a whole book on how our motives often differ from what we say: elephantinthebrain.com But if I have to pick the one area where what we say tends to be the most different from what is really going on, I'll pick: governance.

@ykssaspassky Most everyone claims to be against bias. So you'll have to coordinate around some more specific way to cut it.

@robinhanson Robin, how to re-invent society based on un-biased thinking? Like is there a cultural term for this redo or would it be something fundamentally new?

"Public good" is a good example of a concept that seemed simple, clean and with real-world plausibility, but turned out to be much less relevant and realistic than it seemed.

@StefanFSchubert @GurReptsSohn Compared to most intellectuals, I'm pretty theoretical. But yes, not compared to all.

@GurReptsSohn @robinhanson

One way I see myself as like a philosopher (even if philosophers don't agree), is that I'm pretty unhappy when people talk about important topics using sloppy concepts. Please folks, hold yourself to higher standards re the concepts you use!

While thinkers usually don’t want to directly admit that they seek to impress, sell, or support, we don’t actually much mind observers inferring such motives in us. Few actually much respect those who try to figure stuff out together. overcomingbias.com/p/figure-stuff…

Yes, sloppy concepts can support a useful first cut analysis. But when that gets messy, it might well be time to raise your concept standards.

Yes I'm subtweeting a talk I'm hearing. No I won't publicly say which one.