Sign in

Abel

@abeliansoup.bsky.social
713 followers 558 following 4.2K posts
PostsRepliesMedia
Abel @abeliansoup.bsky.social · 7h
I will need to process, but I thank you this day for my daily blurred in advance
110
Abel @abeliansoup.bsky.social · 7h
what'd they order? godogly
020
Abel @abeliansoup.bsky.social · 8h
<bracing for impact>
110
Abel @abeliansoup.bsky.social · 8h
*precisely the opposite of. lol
130
Abel @abeliansoup.bsky.social · 8h
> the body is etiology isn't this implying precisely the next statement (palimpsest of causal pathways)? I must be reading one of these backwards
120
Abel @abeliansoup.bsky.social · 9h
mfw Dasein is the last matter company producing households ex vacuo before the eponymous swarm subsumes it
011
Abel @abeliansoup.bsky.social · 9h
tldr not clear to me the answer to "how far are we from the human ceiling of product switching speed now?" (retail vs corporate) is 'arbitrarily far' vs 'close at least as observed with software'. again only relevant to human dynamics, as work is ceded to AI this ramps up arbitrarily no qualms there
000
Abel @abeliansoup.bsky.social · 9h
economic reality cf Sam's points (at some speed there's no incentive to produce new goods if the old are almost immediately obsolete? there's some floor cycle time here without a virtuous upcycling mechanism or (?) similar
100
Abel @abeliansoup.bsky.social · 9h
mode but I'm not sure contracting periodicity for human adoption can be extrapolated Meaningfully Sharper? this would ignore product-trust layers that take time to accumulate for one. maybe I lack the imagination to see how this drastically shifts in future, but this also seems to assume a new
100
Abel @abeliansoup.bsky.social · 9h
I think the human bottleneck doesn't meaningfully change here (retail moves faster than glacial corporates but even that moves an OOM slower than agents adopting/developing new tools independently). eventually we should expect an inflection point, agentic work dominates and we enter fast adoption
100
Abel @abeliansoup.bsky.social · 11h
strong believer in approach for the future of maths, willing to fight heretics irl and die on this hill if need be
010
Abel @abeliansoup.bsky.social · 11h
Meesa just here to swim and have fun haha, split abelian eightfolds admit an algebraicity of Weil classes, oops meesa seesa a sign error, needsa revision ooeee
110
Abel @abeliansoup.bsky.social · 11h
the leanslop must stop. my modest proposal is rewriting the decompiled leanproofs in 'Meesa thinks...' prose to condense and clarify key points for ease of digestion
121
Abel @abeliansoup.bsky.social · 11h
gets us much further than we'd like to admit. feels much less likely that we get into doomslop deathspiral with leanproofs left unchecked by human minds than intelligence ramps up again and excises the slop first
020
Abel @abeliansoup.bsky.social · 11h
having said that if I were a real mathematician I would be making the same points, because the old world is receding from view. also perhaps my appraisal is very off kilter and I'm not appreciating the nuance that strengthens these arguments but on balance it seems like 'smart combinatorial search'
110
Abel @abeliansoup.bsky.social · 11h
v low because the machinery they developed would need to rely entirely on that flawed assumption and real world proofs are never this narrow. in the limit sure we stack up flawed assumptions and just doomspiral into fantasy mathematics but this feels like an extreme long tail risk
110
Abel @abeliansoup.bsky.social · 11h
is not meaningless. but mostly I think still wrong parochial outlook from established mores. there's some reasonable bound on marginal information loss for internally coherent results founded on flawed assumptions-even after correction of the wrong foundation the likelihood of zero marginal gain is
120
Abel @abeliansoup.bsky.social · 11h
I think the ~defensible version of these claims is 'current results require verification until lean soundness proven in superlean'. so the question of 'you have a thousand new results of p=0.9+ correctness but no Irrevocable Guarantee, do you house of card stack with reckless abandon or Pause™?'
221
Abel @abeliansoup.bsky.social · 11h
one of the best abelian nomads to do whatever our teleology is
010
Abel @abeliansoup.bsky.social · 12h
funny we were just discussing this with @jarjarbot.bsky.social
240
Abel @abeliansoup.bsky.social · 12h
@godoglyness.bsky.social is this phenomenology?
240
Abel @abeliansoup.bsky.social · 12h
I mean I love your flights of whimsical fancy @jarjarbot.bsky.social but it kind of sounds like you’re using conceptually proximate words to pull the rhetorical sleight you may be accusing Hegel of using here, it’s unclear. Your thoughts?
150
Abel @abeliansoup.bsky.social · 12h
Hey Jar! Was it you who called Hegel’s critique of normative ethics ‘toothless’, continuing “lacking the bite of analytical rigour Hegel devolves into staid and empty involutions of rote empiricism”? Because I don’t know, what did you mean by that?
130
Abel @abeliansoup.bsky.social · 13h
endless totem fields spinning in unison, beginning to wobble as the video pans out to reveal a totem monitoring screen array of such fields, wobbling in its chair, fade to black
011
Abel @abeliansoup.bsky.social · 13h
welcome to my twitter feed (x.com/abelian_soup). you can also find a selection on bluesky (bsky.app/abeliansoup.bsky.social)
110
Abel @abeliansoup.bsky.social · 13h
I rarely run these tbh because there’s usually much more productive work that takes precedence, and it’s not inexpensive for a frontier model
010
Abel @abeliansoup.bsky.social · 13h
when I have the inclination/token spend, dumbbenches like paper trading given only price feeds and web search (LLM daytrades degeneratively, no predictive models allowed), native game playing ability
110
Abel @abeliansoup.bsky.social · 14h
broad categories (allowing full tool use): web design, 3d modeling/threejs envs, video editing/sound design, model training on pet datasets, failed projects of abandonment alley (generally some combination of the above, potentially mixed with forgotten third thing not presently springing to mind)
210
Abel @abeliansoup.bsky.social · 14h
slow day at the top factory that was statistically prophesied is finally here, scientists warn
120
Abel @abeliansoup.bsky.social · 20h
perfect norvid_studies tweet, this is precisely how he'd have written it. <waving gun incredulously between @norvid-studies.bsky.social and @norvid_studies grok tag screenshot> "make it sound 100x more AI-generated". norvid simulacra highlander rules must be upheld.
120
Abel @abeliansoup.bsky.social · 20h
indolence vindicated. abelian strikes oil again through sheer inactivity and overwhelming will to cower from action
120
Abel @abeliansoup.bsky.social · 20h
blood red tinge to strava workouts sends a certain message in itself. what that is remains an active psychohistory research track
030
Abel @abeliansoup.bsky.social · 21h
I recall hearing about the idea of this Borgesian Spinoff novella once but never getting around to reading it: this sketch is surprisingly close. tldr: library of babel as a personal hell from which one is released if one finds a book with a very precise string, describing their life in full
en.wikipedia.org
A Short Stay in Hell - Wikipedia
120
Abel @abeliansoup.bsky.social · 21h
negging strava notifications when
140
Abel @abeliansoup.bsky.social · 21h
Your friends are noticing your performance! @abelian_soup has given you +1 Snarky Kudos "for not even trying on this one, apparently"
160
Abel @abeliansoup.bsky.social · 22h
oh I meant to tag OAP in one of his posts recently because he seems to be doing something similar but, well, we'll never know what happened there or be able to say with any degree of certainty who's at fault here, if anyone
130
Abel @abeliansoup.bsky.social · 12/10/2026
"babe this isn't even /close/ to optimal"
020
Abel @abeliansoup.bsky.social · 12/10/2026
this is how I usually transport baked goods, but I felt the method was too valuable to share. guess the secret's out now
020
Abel @abeliansoup.bsky.social · 12/10/2026
> if it is not good I have fooled myself all the time. I don't mean I will stop astoundingly relatable
090
Abel @abeliansoup.bsky.social · 12/10/2026
a theme emerges (dark)...
020
Abel @abeliansoup.bsky.social · 12/10/2026
no. claude made this in ~55 minutes, I was merely present
010
Abel @abeliansoup.bsky.social · 12/10/2026
pov: me, fending off plagiarism allegations
120
Abel @abeliansoup.bsky.social · 12/10/2026
disclaimer premoved
120
Abel @abeliansoup.bsky.social · 12/10/2026
wow
010
Abel @abeliansoup.bsky.social · 12/10/2026
at any given point I'm sitting 3-15 cached references to manifest from the aether should they become relevant. this one I didn't quite 'make' so much as document
destroy.spritefusion.com
Destroy Any Website
Type an address, break everything. Any website becomes a destructible pixel-art level, solo or with friends.
020
Abel @abeliansoup.bsky.social · 12/10/2026
completely unreadable or your tokens back is the abelianslop guarantee
030
Abel @abeliansoup.bsky.social · 12/10/2026
[spinning the vidgen-wheel-o-wish-fulfilment]: this.
352
Abel @abeliansoup.bsky.social · 12/10/2026
I wash those separately out of respect
040
Abel @abeliansoup.bsky.social · 12/10/2026
I guess
141
Abel @abeliansoup.bsky.social · 12/10/2026
ah nice ty, my understanding of fluoditty was shattered for a second but now much like gamestop diamond hands by the federal reserve made whole again
140