Vecr

15 ms·

241.
by Vecr 2y ago
On the financial level, the state might be too much risk. Personally I would consider if I really needed to have anything to do with Florida at all. If you think that makes sense, consider getting everything out before it gets priced in (
242.
by Vecr 2y ago
I'm looking at a scenario for the counties: Citrus Hardee Hernando Hillsborough Manatee Pasco Pinellas Polk Sumter (Project Phoenix Scenario) Barrier islands are quickly overtopped, provide ~no protection. Water up to 26 feet above tid
243.
by Vecr 2y ago
"Almost unthinkable" according to the Project Phoenix report.
244.
by Vecr 2y ago
Hah, that's the outfit Yudkowsky endorsed. You think he's going to retract? Because almost anyone who knows system administration and LLMs would have told this guy it was a horrible idea.
245.
by Vecr 2y ago
Absolutely it's an argument against God that almost none of those explanations contain God. If you want to enter more evidence or argue priors you can, but in the argument's current state it's highly effective.
246.
by Vecr 2y ago
Obviously that can't be done absolutely, but 1) theoretically there are a huge number of explanations that fit the evidence, possibly an infinite number 2) almost none of them will contain your God Done. You can argue about what system
247.
by Vecr 2y ago
Yes, well. At this point I'm mostly convinced by multiple high quality arguments that Roko's Basilisk is non-functional against any reasonable theory of rationality.
248.
by Vecr 2y ago
Assuming it only as part of that hypothesis, not assuming it for all reality. If it's not impossible, since we can't observe the evidence, its only detriment is complexity by either hypercomputational Simplicity priors (if we don&
249.
by Vecr 2y ago
Not exactly, my argument there was that there would be a huge number of individually more likely (due to being simpler) candidate objects for your criteria, that in combination entirely swamp every conception of God that has ever been writt
250.
by Vecr 2y ago
1) We exist (or at least you do, Cogito, ergo sum, "I think, therefore I am") 2) Since we exist, something had to have lead to us 3) The thing that lead to us was either lead to in turn, or "just existed" 4) If you don&#
251.
by Vecr 2y ago
No. Or, maybe, but go demonstrate a merely gas-based Laplace demon first.
252.
by Vecr 2y ago
Yes? But I assume you aren't arguing for God-as-in-any-object-or-idea-that-meets-a-list-of-criteria, I assume you're arguing for God . God that's described in any book I've read appears to have a lot of complexity, an
253.
by Vecr 2y ago
I don't think your argument follows. Assume there's an uncreated universe that is very suited to running simulations. Assume someone in that universe creates our universe, fully contained inside theirs. Yes, that person is outside
254.
by Vecr 2y ago
You deleted the other comment I was trying to respond to, but I'm interested in a response to the idea that an "object" that meets all the requirements you've given could be simpler than God, and therefore more likely by
255.
by Vecr 2y ago
Maybe "alternative explanations" is better wording. If you don't think God fits into a simplicity detector, could a simplicity detector of some sort that's external to God and every other member of the set of God-like ob
256.
by Vecr 2y ago
I'm not arguing against a single "actual" cause, just that you don't know what the "actual" cause is to the extent that you can exclude all but exactly one explanation. Multiple hypotheses fit the data (or in t
257.
by Vecr 2y ago
Only if you don't think there's there's a singularity (or singularity-replacement) in the future. Maybe against Wheeler? I can't remember. But anyway, if there's a singularity (or suitable singularity-replacement) i
258.
by Vecr 2y ago
Okay, accepting these new criteria (actually somewhat reasonable, if you accept the universe has a creator), intentionality restricts the objects that are acceptable, but if you accept extended Occam's priors that have the capability t
259.
by Vecr 2y ago
No, by the Principle of Multiple Explanations, every "object" that satisfies your conditions is equally viable, everything being equal. You can slap a hypercomputational simplicity prior on that if you want, but that probably mili
260.
by Vecr 2y ago
Well, if you accept the argument it only gets you to an infinite or at least very large number of objects that meet the stated requirements. Not one object in particular. And here "object" is quite vague, I would assume almost all
261.
by Vecr 2y ago
Maybe you're thinking of some strange version of the Omega Point idea? https://en.wikipedia.org/wiki/Omega_Point
262.
by Vecr 2y ago
Roko's basilisk isn't retrocausal, maybe you're thinking of some Nick Land thing? But even there it's hard to tell how literal he is being with the Terminator analogies (if you haven't seen the movie it has time tra
263.
by Vecr 2y ago
Not really, many AI doomer types use AI for various tasks. Obviously they need to for research, but even outside that they say they get benefit.
264.
by Vecr 2y ago
If you claim something is an "affront to humanity" it clearly induces doubt in your opponent that you would use it. Not a good negotiating position where deterrence is concerned.
265.
by Vecr 2y ago
They really need to use positive-pressure helmets or full face masks of some sort. SCBA or PAPR.
266.
by Vecr 2y ago
Won't the scientist use some relatively secure/private model to fraud-check their own work before submitting? If it catches something, they would just improve the fraud.
267.
by Vecr 2y ago
Rust's Rayon could probably have a mode with additional heuristics, so it wouldn't multi-thread if it guessed the memory usage increase or overhead would be too much. Not really automatic, but for some iterator chains you can just
268.
by Vecr 2y ago
No it's not, for example Yudkowsky gives his standard talk on there. I think it matters if the guest has something ready to go for the allotted time.
269.
by Vecr 2y ago
Hope he has something in place for the kidnap risk. Also hope his lawyers credibly claim they will demand continual proof-of-life to engage, otherwise kidnapping might not be the end to it.
270.
by Vecr 2y ago
Are they limiting the input to a 1% cap, then multiplying the difference from the actual model by 0.000000001? If so it might not be too bad in practice, but it's still bad science. It's like someone today took planetary epicycles

More ›