Opens profile photo
Follow
Nate Soares
@So8res
communication handbook: tinyurl.com/yn32dsx7
Berkeleymindingourway.comJoined January 2011

Nate Soares’s Tweets

(Some aliens might be 10-fingered! AIs are less likely to be 10-fingered, or to even have fingers in the relevant sense, in lieu of specific effort in that regard! See also: twitter.com/ESYudkowsky/st)
Quote Tweet
Why? Obviously not because silicon can't implement kindness; of course it can. Obviously not because it's impossible to blunder into niceness by accident; if so, I wouldn't expect it about 5% of aliens. Rather it's that - on my model - kindness is 5% dense in one particular… Show more
Show this thread
15
Show this thread
I answer: no; the symmetry is that other minds have other ends that their intelligence reinforces; kindness is not priviledged in cognition any more than Earth was priviledged as the center of the universe; imagining all minds as kind is like imagining all aliens as 10-fingered.
2
23
Show this thread
I sometimes hear "But isn't it sheer hubris to believe that humans are rare with the property that they become more kind and compassionate as they become more intelligent and mature? Isn't that akin to believing we're at the center of the universe?"
2
17
Show this thread
The generalized lesson from geocentrism being false is that the laws of physics don't particularly care about us. It's not that everywhere must be similar to here along the axes that are particularly salient to us.
1
18
Show this thread
I've encountered some confusion about which direction "geocentrism was false" generalizes. Correct use: "Earth probably isn't at the center of the universe". Incorrect use: "All aliens probably have two arms with five fingers each."
1
43
Show this thread
Abolish the FDA; repeal all occupational licensing including in healthcare; decriminalize sex work; decriminalize housing; delegalize marriage and college; allow free banking. Regulate nuclear materials. Ban >GPT4 model training runs and gain-of-function research on pathogens.
165
573
Show this thread
Okay, let’s go through these one by one. I’m glad Yann is engaging here, and further I think his proposal has some value for alignment (& I hope Meta AI pursues it instead of simply scaling AR-LLMs). With that said, I think his points come wildly short from proving his conclusion
Quote Tweet
Replying to @ESYudkowsky
1. Since nobody knows how to make an AGI, it is tautological that nobody knows how to make a nonbad AGI. 2. That said, if the architecture of the system is such that its outputs *must* minimize a set of objective functions at *inference* *time* , then the problem is merely to… Show more
5
42
Show this thread
(ftr: I signed onto futureoflife.org/open-letter/pa because I think that the current path leads to destruction, and that the letter's suggestions are marginal steps in the right direction, not because I endorse all its arguments, nor because I think those steps would help all that much.)
7
185
Show this thread
(no disrespect to ppl who live in villages; i don't think y'all are the best when it comes to dominating planets or whatever, but in my book that's a virtue)
16
Show this thread
Also, academia isn't reliably good at statistics, nor at analyzing social/psychological phenomena, so it's a strange point of comparison. Unless you're just swatting down someone who you think has risen above their station, ofc, in which case the objections make perfect sense.
3
69
Show this thread
Also, while I'm on the topic: a fun hidden fact about Earth is that you don't actually need a license to collect and analyze data! No matter what the "do you have a degree" gatekeepers insinuate.
1
95
Show this thread
For people who present as caring a bunch about data integrity, they're weirdly unresponsive to the data on their pet theory that Aella's polling population differs radically from a bigger and more diverse survey population. (The data isn't kind to their theory.)
2
93
Show this thread
But it's even worse than that! Because while Aella is mainly reporting results from big diverse surveys (while treating her polls as mere recreation), we can compare her big surveys to her polls and see that the polls are actually pretty dang accurate.
3
44
Show this thread
A common misconception of Aella's research is that it's constructed from Twitter-polls of her followers. Nope! When she reports research results, she's talking about huge surveys of fairly diverse populations. (Much bigger and more diverse than is usual in academia!)
Quote Tweet
People often say that my research is "twitter polls." I do a ton of twitter polls, but I primarily use them to gauge what might be potentially interesting topics for more thorough surveys in the future! My actual research is stuff like this: aella.substack.com/p/who-took-my-
5
155
Show this thread
(But "society is net-worse at responding to the second event because people're exhausted and pained from the first one" is a way reality can go, despite all your rose-tinted hopes about people suddenly waking up to the dangers in the face of a warning shot.)
2
34
Show this thread
(To be clear, I expect some pandemic responses to be better and some to be worse in the wake of COVID, and would be interested to learn what distinguishes them--although hopefully I won't get the opportunity, ofc.)
1
20
Show this thread
I'm grimly amused that Earth seems perhaps "burned out" about pandemics; seems perhaps *less* likely to react quickly and competently than pre-COVID. (Which does not bode well for the "surely humanity will get its act together after a warning shot" theory of AI alignment.)
Quote Tweet
Can some people either start betting this market down or start panicking please? manifold.markets/NathanpmYoung/
Show this thread
Image
6
87
Show this thread
It's like being in a room full of LEGO machines, and you look at the machine that reads instructions and assembles the other machines, and it's built not out of LEGO but out of cleverly contorted instruction booklets.
3
113
Show this thread
3. a community is probably stronger when its members just blurt out their beliefs (while meticulously being kind to each other). it's much easier to lose your way if you live in a mental world where PR is king over honesty and integrity. HT
2
106
Show this thread
2. people and institutions lauded as genius are often held together by only bubble-gum, wishes, and a favorable environment. if you rely on those people/institutions to accomplish great feats of competence under pressure, you're in touble.
3
77
Show this thread
1. take note: the status quo is not always stable. your "criticism contest" can fail to reveal the giant risk that was staring you in the face the whole time. regularities you were implicitly and unconsciously relying on can evaporate overnight.
2
68
Show this thread
honestly the best argument for most school im aware of rn is that it functions as a babysitting place for parents who need to work but we should at least admit that and optimize *for* that, instead of systematically destroying our youth's relationship to learning
11
189
Show this thread
the "curse of cryonics" is when a problem is both weird and very important, but it's sitting right next to other weird problems that are even more important, so everyone who's able to notice weird problems works on something else instead.
11
94
And don't worry about loss of intellectual diversity; I guarantee that if you wholeheartedly adopt all the correct views you can find then you'll end up with your own unique viewpoint before long.
49
Show this thread
and people assure me that governments will start acting sane and reasonable around AI in the wake of "warning shot" accidents
Quote Tweet
Peter Daszak has received another grant from the NIH… …to study bat coronaviruses in the wild. After everything the world has just been though. After all the risky research that was supposed to protect us from a global pandemic failed to stop one.
Show this thread
9
206
A recurring special case is with jargon. Someone says, here's a concept, a new term for it, and some writing to connect it to other things. Objection! You should use an older, very slightly mismatched term. It has four extra syllables, no mnemonic, and incorrect connotations.
Quote Tweet
it's even worse than @yashkaf depicts: big progress often comes from lots of small reconceptualizations. the "i can't distinguish your idea from a worse one in the literature" police are punishing real progress. twitter.com/yashkaf/status…
2
77
it's even worse than depicts: big progress often comes from lots of small reconceptualizations. the "i can't distinguish your idea from a worse one in the literature" police are punishing real progress.
Quote Tweet
this is the best part of TPOT the fuck do I care that someone 200 years ago had the same realization and wrote it down somewhere? fucking good for them! what does it matter if I read it in a thread or a book or a thread quoting the book if it's the same idea? twitter.com/profectedroom/…
Show this thread
6
198
ok my new theory is that girls can both have a feeling active *and* be forming words at the same time. and this is just, like, how they live their lives
18
84