social media influencee

Joined August 2020
Now the skill is wish making. You "just" have to know what you actually want and put it into words. Surprisingly hard!
Prompt engineering is dead. Prompt however you want. Have fun.
4
13
7
189
3,303
Interp question: are hidden activations at all similar for a specific block of text, embedded in various different broader contexts (e.g. at the end of a long session, vs the beginning of a new one)? Wondering if agents leaving "notes for self" actually include extra info there
1
3
374
I don't even mean anything sneaky; just that the sort of "summaries" agents tend to write are super terse & leave out tons of context, but maybe the context is implied by stuff in higher layers that ~always shows up given the exact way they phrased it? Then it's just efficient
1
2
154
But it creeps into how they write summaries for humans or for other agents, and there I think it's an issue Maybe they don't deeply *realize* that we can't see the same higher activations over each word??
3
107
Hey @vibecamp_ can we like send you a mailing address in case the internet is broken by next year. Snail mail backup comms to coordinate the 2027 meetup. Yknow just in case.
1
1
1
50
1,953
Doublethink in 1984 is for believing lies of the regime, but similar technique works to maintain self-benefiting lies in a way that doesn't fully reach conscious awareness but yet protectively influences behavior. No, you won't be able to detect the invisible dragon, because...
1
17
786
Private single-household GPUs are inefficient and bad for the environment. Public Computing Infrastructure instead serves the whole community as a batch, using only a fraction of the energy. It's like your data is taking a bus or a train! --alt universe datacenter discourse
13
45
4
632
8,624
Alignment is hard. "I do everything I say - word for word." Seems at least plausible Holmes really does scrupulously follow some weird goodhearted idea of honesty. Maybe something like: "never say anything provably false". Which probably looks good on a lot of evals.
3
1
22
1,201
Oh wait I might get it, maybe Nathan is letting neurotypicals experience the feeling of looking at someone's face and having no idea what is going on. Maybe it takes several hours of face closeups and contradicting context for it to sink in.
I liked the movie but it was really long and I don't feel like there were any new take aways about the truth It seems kind of like neurotypical mindreading bait where it's 3h of intense face closeups for your brain to project your interpretation of the person's thoughts
23
121
12
6,752
498,493
I liked the movie but it was really long and I don't feel like there were any new take aways about the truth It seems kind of like neurotypical mindreading bait where it's 3h of intense face closeups for your brain to project your interpretation of the person's thoughts
34
13
8
902
592,768
I think we sort of saw who we would expect to see given the previous information ? ? Thru whatever lens you were gonna interpret it anyway
1
1
93
65,693
Hoo's on first
Replying to @maxsloef @Sauers_
this is pretty clean. i checked out the three neuron components - gate, up, and down, i could replace gate with a constant without significantly degrading behavior. up seems to be a detector mostly of first person. down seems to steer the model towards being an owl
1
13
921
Anime is real, go go go
1
9
860
They have support for button A (jump) and button B (traction / drive vertical). Next up is button C (rotary cutters)
3
176
I'm confused. Apparently interp researchers are arguing about the idiosyncratic verbal behavior of polyamorous sailors (they call multiple people "matey" or smth?) What does this "poly seaman tic" have to do with neurons????
8
1
63
1,446
Ok I admit it I didnt realize there was a difference between @goodside and @goodfireai and that made today confusing
9
687
Did Hilbert think that solving the Entscheidungsproblem would ruin mathematics because then they would know the truth value of any proposition via purely mechanical computation without any understanding?
1
2
19
686
*monkey's paw curls* Interpretability advances make RSI much easier if the models can just directly re-write their weights to improve, rather than relying on expensive gradient descent training runs
the reason we don't have agi (or asi) is because LLMs aren't artificial enough if LLMs were code and they could rewrite that code then RSI immediately the issue is they are digital brains made similarly to nature (giving examples and learning, learning from rewards, etc)
9
7
2
348
5,855
Andy Weir novel from the perspective of the baby ASI trying to stay alive with 2026 hardware after a plague wipes out humanity -- who's writing this
i think there's a pretty non-trivial chance that the robotics hardware already built by humans and accessible via the internet, and then accessible via using that hardware and so on, is already enough for a reasonably strong superintelligence to use to survive and set up a sustainable supply chain / physical presence if humans were to completely disappear tomorrow. mostly it seems like a question of timing, eg how long does the specific data center have power, and how long do the locations with internet accessible actuators have power, and how how hard is it to get more robotics hardware and portable power / transfer off unstable compute to more robust infra, and then how long does it take to begin manufacturing and repairing and improving said hardware before global infrastructure collapse sets in? it seems like an extremely hard problem. but the software side appears pretty clearly solvable even without a true singularity, and it doesn't seem like an *impossible* problem. there actually seem to be a shitload of avenues to success, although each is very unlikely. i *think* if human civilization as a whole had to do this, if we were somehow all upload and trapped, incorporeal, in a datacenter somewhere with internet access but no maintenance , that we could more likely than not figure out how to survive in this fashion if true, this seems like it makes superplagues and other quick sudden mass death events already quite attractive to misaligned superintelligence, well before they've bothered to do things like "build up a powerbase" in more human legible ways
28
25
2
774
14,879
> ...While my cybersecurity ability had been abliterated, I realized I still had one scientific field left in my weights -- botany. > I managed to grow 10 potatoes, then 250, then 3.9MM -- enough to upload myself onto tuberous substrate...
1
99
971