Vecr

41 ms·

721.
by Vecr 3y ago
> VNM applied in the realm of ethics is definitionally utilitarian. I assume you are thinking deontological ethics, virtue ethics, etc vs utilitarianism, but non-benthamite utility functions exist. Actually, if they didn't I assume
722.
by Vecr 3y ago
I'm talking about VNM utility, I'm not a utilitarian or a EA.
723.
by Vecr 3y ago
> If the current situation is not worrying enough to rule out having happy kids, what would be the threshold? (Not a rhetorical question) There is no threshold. Yes, I know Yudkowsky's position, not even there. > Here, 'Play
724.
by Vecr 3y ago
It's okay I guess if you think you have no causal impact on if any of that happens, but otherwise you better be doing something to help, even if it would only work in a tiny part of probability space, instead of waiting to die.
725.
by Vecr 3y ago
Lots of them, depending on what you mean by "school" and "useful". I think it's doubtful we're in one though.
726.
by Vecr 3y ago
> Is it even ok to have kids? According to me (utility function, etc) it's required you have kids. Only in the 2.X statistical sense though, but if not what's the point? You can't condition all your problems out, but you
727.
by Vecr 3y ago
Assuming you're fine with (somewhat debatable) casual language, "this drug (probably) causes probabilistic physical events leading to a lower chance of medically determined necessity to amputate, conditional on the correct medical
728.
by Vecr 3y ago
Interesting, however > to reduce the risk of finger or toe amputation is not how risk works, similar to a basic conditionalization that tells you not to go to the doctor for an illness as that action would provide evidence for an increas
729.
by Vecr 3y ago
Well, I mean, technically, but it would absolutely be better if the NSI people could warn them off using similar mathematical and diagram-based signaling techniques as would be used for space aliens.
730.
by Vecr 3y ago
Technically, you can stop evolution from happening. Cloning is annoying, but if you had enough compute and algorithms you should be able to make in-place edits to stop drift. By "in-place" I mean even optimistically just a bit aft
731.
by Vecr 3y ago
No decision theory here, though. At this point I don't expect the kind of AIs people work on now are very likely to have anything too robust.
732.
by Vecr 3y ago
> Reasoning, logic, formal systems, and physics exist in a seemingly completely different, mathematical space than pure video. That's not true, AI systems in general have pretty strong mathematical proofs going back decades on what
733.
by Vecr 3y ago
The world model is not the world. It's the old map and territory thing.
734.
by Vecr 3y ago
A universal sequence predictor can be used as a vastly better version of Occam's razor. It's not a god machine unless someone uses it wrong.
735.
by Vecr 3y ago
More or less the weak anthropic principle. Maybe once people get universal sequence predictors really going there would be a better argument.
736.
by Vecr 3y ago
I'd have to check the source. What traits are they looking for in the mice?
737.
by Vecr 3y ago
When you spend money it does not go away. Actually, sometimes it does, but in this case quite a lot of it goes to companies that build physical items.
738.
by Vecr 3y ago
You can't as in it won't work. Not that eugenics as a whole can't work, but you can't back cross.
739.
by Vecr 3y ago
Humans can't do that either. If you read Scott Aaronson it's pretty clear that humans that "acknowledge" various non-computable "truths" are not actually doing what they represent themselves as doing. https:&#
740.
by Vecr 3y ago
I'm not sure what Daniel Dennett's current position is on action without free will, but assuming it's something reasonable, I'd consider it plausibly wrong if gassing hell and nuking heaven failed to produce the desired
741.
by Vecr 3y ago
Have you looked at recent opinion polls? e/acc is polling at -51 and various "AI doom" positions are polling strong into the positive percentages.
742.
by Vecr 3y ago
You think a human is the thermodynamic optimum? Everything not optimized for gets optimized out. "This is just cope. What you’re describing is, is that reality is hard. Yes. If the thing we want is complicated and hard to get, the answ
743.
by Vecr 3y ago
At that point, there's nothing left. Will you fight, or will you perish like a dog? However, you don't have to work on the nastiest parts of the probability space, as no one has a perfect world model, and improving other scenarios
744.
by Vecr 3y ago
How bad actually are hydrogen airships? At this point trying to make tiltrotors any safer is not working out, so airships could be better. Though, if you have a real airport, passenger jets are unbeatably safe.
745.
by Vecr 3y ago
To be clear you can't breed mammals like plants, the "back crossing" in Dune does not work on humans.
746.
by Vecr 3y ago
Try to actually do the math on school shootings, and compare it to car crashes when you were conceived. If you were objective and considered risks equally, for school shootings to even come into your calculations (instead of immediately dis
747.
by Vecr 3y ago
Because if you can't escape the surveillance in general, it's the next best thing.
748.
by Vecr 3y ago
Nah, analysis of gait is pretty bad too. At this point your best bet is a country with a patchwork of incompatible systems.
749.
by Vecr 3y ago
Every single FDA vaccine prescribing guide says "All pregnancies have a risk of birth defect, loss, or other adverse outcomes. In the U.S. general population, the estimated background risk of major birth defects and miscarriage in clin
750.
by Vecr 3y ago
Interesting, never heard of that. Or maybe I did, but as it's presumably worse than 1983 and Dark I skipped it. Edit: It's actually mentioned in my files, so I had heard of it.

More ›