Joscha Bach

@EmilMieilica Like Starlink, they would be in low orbit and burn up around the time the GPUs become obsolete

@Plinz And I thought it was about the Kessler syndrome

we need to prevent elon from building datacenters in the sky because it will use up all the water in the clouds

@dioscuri It's a matter of taste; without knowing yours it's hard to say. All three are games for autistic people; especially Dwarf fortress asks for a massive time investment. I would recommend Factorio, it's incredibly elegant but has no social sim aspects

Thinking of playing one of these games next. Have never touched any of them before but feel I need to play at least one. Any strong recommendations among them?

@losslandscape We seem to be in the timeline where Elon wins against all odds

@Plinz They're going to stop them from being built in space. There will be some excuse. There is no escape.

@compendiumvitae @brivael There are reasons to believe that our reality is not a simulation, but the three he gives are not them

@brivael @Plinz

"Et si toute votre vie n'était qu'un programme informatique gigantesque ?" Non. Et le raisonnement de Bostrom contient un défaut structurel qu'on ne discute presque jamais. L'argument de la simulation suppose qu'une civilisation suffisamment avancée pourrait "tourner" des univers comme on tourne un programme. Mais cette intuition transporte clandestinement trois hypothèses du XXe siècle informatique qu'elle ne justifie jamais : 1. Que le réel est calculable. Or un calcul opère sur des états discrets, finis, dénombrables. Le réel — ne serait-ce que les champs quantiques d'un univers à la résolution de Planck — exige ~10^123 bits/seconde. Multiplié par toutes les configurations possibles, on dépasse la borne de Bekenstein de tout substrat physique imaginable. Le contenant doit être plus grand que le contenu. Mathématiquement impossible. 2. Que simuler une chose, c'est la rendre réelle. Une simulation parfaite de pluie ne mouille personne. Une simulation parfaite d'un univers est un modèle isomorphe de cet univers — pas cet univers. La carte n'est jamais le territoire, même quand elle est parfaite. Bostrom confond représentation et instanciation. 3. Que la conscience émerge de la structure causale, indépendamment du substrat. C'est le fonctionnalisme naïf hérité de Putnam. Mais rien ne prouve qu'un pattern de transistors instancie une expérience subjective. C'est une croyance, pas un théorème. Le vrai retournement : si tu crois que nous vivons dans une simulation, tu crois que la réalité est faite de code. Si tu refuses cette idée, tu dois assumer que la réalité a une épaisseur que le calcul ne capturera jamais. C'est la même intuition à toutes les échelles. Un dashboard n'est pas une boîte. Un avatar IA n'est pas une personne. Un modèle économique n'est pas un marché. Le réel déborde toujours sa représentation — c'est précisément ce qui fait qu'il est le réel et pas son image. Bostrom est un intellectuel brillant qui modélise. Mais quand tu construis vraiment des choses dans le monde, tu apprends vite que le monde gagne toujours contre le modèle. Le territoire mange la carte. La masterclass L'hypothèse de la simulation est devenue la religion de la Silicon Valley parce qu'elle flatte exactement ce que les ingénieurs veulent croire : que tout est code, que tout est descriptible, que tout est en dernière instance manipulable depuis une console. C'est une eschatologie déguisée en philosophie analytique. Mais la vraie question n'est pas "vivons-nous dans une simulation ?" — c'est "pourquoi a-t-on tellement envie d'y croire ?" Réponse : parce que ça rend le monde contrôlable. Si c'est du code, on peut le hacker. Si c'est un programme, il y a un programmeur — quelqu'un en charge, un sens, une sortie possible. C'est du gnosticisme high-tech : le monde matériel est une illusion, la vraie réalité est ailleurs, le salut viendra par la connaissance technique. Le builder authentique fait l'inverse. Il accepte que le monde a une opacité irréductible. Il construit dans le réel, pas au-dessus de lui. Il sait que sa simulation, son modèle, son agent, son produit, ne sont pas la chose — ce sont des outils pour interagir avec la chose. C'est pour ça qu'on construit. Parce qu'on a compris que la réalité ne se contemple pas depuis l'extérieur — elle ne peut pas l'être, il n'y a pas d'extérieur. On y est jusqu'au cou. Et la seule réponse digne, c'est d'agir dedans. Bostrom rêve d'une console. Nous, on tape sur le clavier du monde réel.

@JosephJacks_ oh no your account has been taken over by a slop generator

In 1964, Stanislaw Ulam posed a question that has haunted computer science ever since: what is the optimal physical substrate for computation? Not a faster chip. Not a better architecture. The substrate itself — the material. Ulam envisioned a hypothetical substance in which every atom participates in useful information processing. No wasted mass. No passive scaffolding. No distinction between the computer and the thing it is made of. The theoretical physics community gave this substance a name — computronium — and then spent sixty years failing to build it. Silicon wastes most of its atomic structure on lattice stability. Gallium arsenide, graphene, even photonic crystals — every engineered material reserves the vast majority of its mass for structural integrity while only a thin surface layer does the actual computing. But > 500 million years ago, in water, at room temperature, a single 110-kiloDalton protein solved Ulam’s problem… 💧 Tubulin — one heterodimer, one polymerization grammar — builds a whip that beats mucus from your lungs, a bipolar spindle that segregates your chromosomes to the nanometer, a helical cage that resists the compression of every heartbeat, a crystalline rod that propagates a sine wave with no motor protein, a cone-shaped drill made of aberrant protofilaments that pierces a host cell membrane, a geodesic starburst that templates mineral silica into the most symmetric structures in the natural world, a temporary scaffolding skirt that sculpts a sperm head into its species-specific shape and then destroys itself, and a row of tilted cilia in an embryonic node whose forty-degree angle of inclination determines whether your heart develops on the left side of your body or the right. Fifteen + architectures. One molecule. No engineered material on Earth (yet!) can reconfigure its own geometry at room temperature in aqueous solution to perform this range of mechanical, structural, sensory, and information-processing tasks. This is what Ulam was looking for. Not a material that computes upon a structure — a material whose structure is the computation. Every polymerization event is a decision. Every lattice geometry is an output. The microtubule morphology IS the computation.

@Romy_Holland en.wikipedia.org/wiki/Swatch_In…

it’s weird that every culture uses the same timekeeping. i know there have been other calendar systems, but everyone settled on 24 hour days and 60 minute hours and 60 second minutes. these are weird numbers! why didn’t anyone implement metric timekeeping?

@howdidyouseeme1 @tszzl Yes, he instantly recognized who the incompetent clown is; his entire disagreement is about the fact that one of his students once tried to train an rnn

@Plinz @tszzl But it is actually true that the post was about him and was understood as such by anyone who knew the context. Threatening to sue because someone said you've never trained a neural net is absurd of course but the part about him being the target of the post is not.

gary marcus, beyond just being plainly & clownishly wrong for years, is a slimy and bad person and you shouldn’t encourage him even as a rhetorical object

@SimoneSyed @202accepted I knew it. Kidney stones come from eating too much vegetables and underprocessed food

@202accepted I had an ex who had one - it was the most pain I had ever seen him in. It's a stone, multiple stones often. One way they form is consuming too much oxalates in like spinach and kale and not enough calcium... and not staying hydrated. I'm sure there's also a genetic component

how do you get a kidney stone? no-one in my entire genetic line has any clue what this is it is literally not a thing we have ever encountered, like nobody in the entire extended family has a frame of reference for this it’s borderline like we think it’s maybe fake news

@lumpenspace @OyeBhartsingh Yes, Indians make excellent Americans

@Plinz @OyeBhartsingh there really must be too many h1bs then

Why is USA most hated country in India?

My wife discovered this years ago.

@LughSpear @kanair All causal structure requires dynamics, including consciousness. If something does not move, it does not do anything. That means you need at least state transitions. (Time can be understood as state transitions relative to the clock of its observer.)

@Plinz @kanair Side question : is consciousness linked to time ? how can you imagine minds without time, with or without physical substrate ?

Philosophical zombies are a being physically identical to us, but without consciousness. But what about the opposite “philosophical ghost”? Pure consciousness with no physical substrate. Strangely, this feels much less conceivable.

@PNASNews Congratulations, Lee!

One of the most-viewed PNAS articles in the last week is “Chemputer and chemputation—A universal chemical compound synthesis machine.” Explore the article here: ow.ly/V53j50YKzbw. For more trending articles, visit ow.ly/2Mfa50YKzbA.

I don't believe that mind is fundamental, because it is complex (whereas matter is simple) and varies with changes in its physical substrate. But if we can think of the physical universe as a mathematical object that produces causal dynamics without having a substrate itself (because base reality has per definition nothing "below" itself) , it does not seem to be harder to think of a mind that is similarly a mathematical object without further substrate layers below it. This appears to defeat your intuition, no?

@Plinz It’s hard to imagine, but it’s interesting to think what it means.

@DavidBanik @StuartHameroff Some folks believe that even individual cells are conscious, some think it requires multicellularity, some think it requires mammalian brains etc. Most researchers think that collapse theories don't provide plausible causal mechanisms for generating consciousness.

@Plinz @StuartHameroff Potatoes lack a brain, which is required for generating the macroscopic quantum states whose collapse results in conscious states.

Are you saying water is conscious? Is everything conscious? What’s your point?

@MicaelMarch @kanair A lot of people are arguing that LLMs are philosophical zombies

@kanair There are no philosophical zombies. They are a red herring / a consequence of wrong or poor epistemological thought. The alternative you mention, though; I see nothing wrong with it.