Sabitlenmiş Tweet
it has been a little bit more than three years since i first started writing on smoothbrains.net. i would like to pause blogging for now! here is a retrospective to cap things off: smoothbrains.net/posts/2025-10-…
English
Cube Flipper
9.5K posts
Cube Flipper
@cube_flipper
As a human being, the kindest thing you can do to your brain is to not think.
Katılım Ağustos 2022
1.4K Takip Edilen4.3K Takipçiler
@cube_flipper I would guess it's probably the removal of V1 gamma that causes the blindness, so that part of the field is just noise and has no signal in a fi ks theory
English
This talk is very good, but with blind sight I am always taken by the assumption that the blind functionality is non conscious and that V1 is the source of visual consciousness
Sichu Lu@lu_sichu
English
@Qivshi1 i think at this stage i model ephaptic coupling as playing a much stronger role early in development, to guide neuronal growth/global coherency over a larger scale than would otherwise be possible. those neurons get pretty good at emulating this over time (and this saves energy)
English
@Qivshi1 could be something like ephaptic coupling or whatever is responsible for global phenomenal coherence is disrupted, but the neurons are already good at modelling the field's behaviour
English
@cxgonzalez @morallawwithin "i don't think that could happen" @morallawwithin if you deny that direct reports are viable evidence, you are going to have to investigate this yourself
English
listening, "i don't even know what a preference is" lmao i am so with you. that said i think we will likely have to formalise/construct a definition of "preferences" in phenomenological terms (the correct terms) in order to eventually convince people that preferences are based in something more universal
English
me: what if instead of grounding morality in preferences/desires, we ground in Valence Realism, where Good and Bad are grounded in the valence of consciousness itself?
@morallawwithin: this gets the order of explanation backwards. experiences only present as Good if we prefer them
me: preferences don’t seem to be fundamental, they seem to be overlaid on top of direct experience, valence seems prior to preferences in non-dual awareness
florence: but we only have evidence of these states of consciousness after the fact, once we’ve rebooted our conceptual scaffolding
English
@cxgonzalez @IntractableLion some guy on reddit getting mad about it a long time ago reddit.com/r/badscience/c…
English
@cube_flipper @IntractableLion oh you don't like the explanations people offer alongside the picture, that makes sense
English
Cube Flipper retweetledi
"It's because of quantum mechanics" is one of those non-answers that sounds good to laymen without informing, drives me nuts. The real answer is that the mathematics of space filling curves are universal regardless of scale, from bones to galactic superstructures to cabbages
Brian Cox@ProfBrianCox
This was a very good question from Ryan - why does the Universe on the largest scales resemble the Universe on the smallest scales? The answer, for the case of the distribution of galaxies on the sky, is that the pattern we see has its origin in quantum mechanical fluctuations during inflation - before the hot big bang (at least according to inflationary cosmology, which is the most widely accepted model at present). Hail Mary is a very good film by the way :-)
English
@cxgonzalez @IntractableLion op is a great explanation, i am agreeing with it. the picture has been floating around the internet for aeons and is frequently used to justify a variety of wook science nonsense that is not the sane explanation given by op
English
@cube_flipper @IntractableLion why does that picture bother you but the explanation is a good one? seems like the picture is good at pointing out there are scale-agnostic laws and OP is a good explanation for that
English
English
@algekalipso @liminal_warmth Salvia Does Not Like To Be Smoked and will Punish You for doing so
English
I don’t understand how anyone could hear about salvia trip reports like “I spent 5 years in the bug dimension” or “I was a sentient inanimate hose in my back yard for 20 years” and still be like “well maybe I’ll just try it once…”
English
@cube_flipper @IntractableLion i’m still not saying we live in a simulation but…x.com/ascetic_shadow…
shadow@ascetic_shadow
listen, i’m not saying we are in a simulation. i’m just saying if i were going to make a simulation i would populate it with elements by making small variations to preexisting patterns to save on compute
English
@joseph_h_garvin @quetzal_rainbow oh i see the point is once they added agentic capabilities it figured it out
English
@cube_flipper @quetzal_rainbow A human trained on all publicly available brainfuck programs would score much better
English
"Language models solve 10% of tasks in languages which would make your brain boil out of eye sockets which prove that they only memorize stuff"
François Chollet@fchollet
This is more evidence that current frontier models remain completely reliant on content-level memorization, as opposed to higher-level generalizable knowledge (such as metalearning knowledge, problem-solving strategies...)
English
@viemccoy depends i guess on the cost of a bullet, which is either zero or infinite
English
@viemccoy it's kind of apples and oranges, i actually expect that defining cognition should be trickier than pinning down a definition of consciousness, but doing so shouldn't require solving any metaphysical questions or at least biting any metaphysical bullets
English
I think cognition research is far more tractable than consciousness research. I actually think consciousness, as we've ill-defined it, lives somewhere near the type-error-tragedy I discuss in Semiotic Triage - attempting to study it with scientific tools might actually be barking up the wrong tree. This is obviously not true for cognition, however, and the mechanisms of thinking, solving problems, and creating knowledge. This is distinct from consciousness which seems to almost plainly be the scientific attempt to integrate a phenomenological soul into an empirical world-model. I am open to this being possible, but I am largely skeptical.
(People have attempted to redefine consciousness to mean basically-cognition, but it hasn't stuck)
English
Cube Flipper retweetledi
anyway this is all relatively fine in the case of language models, where this gets dangerous is when the labs start building new architectures/the perfect puzzle vortex x.com/ChristosTzamos…
Christos Tzamos@ChristosTzamos
1/4 LLMs solve research grade math problems but struggle with basic calculations. We bridge this gap by turning them to computers. We built a computer INSIDE a transformer that can run programs for millions of steps in seconds solving even the hardest Sudokus with 100% accuracy
English
post-training the model to solve math puzzles rather than more intuitionistic, context-dependent problems perhaps also plays a part. blah. think smooth vs. striated state spaces, also – does every part have a chance to influence holistic behaviour? there's a sweet spot here
English
(epistemic status: extremely vibes-based speculation) i think this is plausible as a function of how expanded a language model's habitual attentional modes are. (functional attention, not attention is all you need attention – akin to human attentional patterns).
Brangus🔍⏹️@RatOrthodox
I have heard that some anthropic safety leadership are going around telling people that alignment is a solved problem. This seems like a predictable failure to me, and I would like people who thought that funneling talent towards anthropic was a good idea to think about it.
English