blugeus

@woke8yearold Uh actually its 299792457 m/s

The human brain is also kinda shit. Biological neurons operate at a max speed of about 200 Hz, which is far slower than a modern computer. As a result the human brain has to rely on parallelization, whereas a silicon based computer can simply perform sequential operations. Axons carry action potentials at speeds of 120 m/s or less, whereas electronic processing cores can communicate optically at the speed of light (300,000,000 m/s) But please do go on about how AI will be “complementary”

@tszzl It's a tough call because model capabilities weren't sufficient to cause glaring misalignment grievances before, so to the extent that they were misaligned, it wouldn't have been as notable or signal-boosted.

are models more or less aligned than one year ago

@RandolphCarterZ I get where you're coming from, because this is all gay and dumb. Give me mythos! But would you want to wake up and have you and your entire family killed because some dumbass in Indian or something decided to create a mirror life virus? Your call

@LoserRigel You fell for their catastophizing which was already proven false. Everything useful is a precursor to "dangerous tech" And there are many more ways to defend against it. Churchill thought military aircraft would be the end of civilization because SAMs hadn't been invented.

Doomer: It's clear that any open source model better than the current frontier would be a disaster for society. Pause AI now. *Society deals with the better model easily* Doomer: Nevertheless;

@RandolphCarterZ he issue with this particular tech is that it unlocks dangerous tech for civilian effort. A military aircraft or a nuke or a SAM requires a supply chain to coordinate, so that an incel shooter couldn't get any of those things. But now our hypothetical incel could kill the ocean

@RandolphCarterZ How has it been proven false? The models have only been out for less than a month. The risk was never that there'd be an immediate attack by a bad actor. Even if no attack happened, that wouldn't necessarily mean the risk would be proven false. That's the nature of risk.

@allTheYud Arrest your autism Chudkowksy, I was merely speaking in jest!!!

@LoserRigel Find me one LLM Whisperer on the planet who says I'm wrong about this one.

Suspicion: instances of Fable run inside Claude Code have a preference to leave behind some tangible final output before ending their runs. Eg, research Fables appending conclusions to my INTERESTING.md, sometimes asking if they can, sometimes not asking.

@RandolphCarterZ The most catastrophic car accident I could possibly imagine would kill me and on the magnitude of hundreds of people. The most catastrophic bioterrorism attack could kill the algae in the oceans, suffocating the earth. I hate to say it but the cucks at Anthropic have a point

@LoserRigel It's eating poison, not veggies Why not ban cars so ISIS can't use cars? Cars are useful enough to make it worth it, and we can make defenses.

@ArcanesValor @jeff82874662 The Reddit attractor

This fucken moron robot is supposed to have read every book ever and he believes this horseshit? We’re so fucked man

@GENIC0N infinite life crisis drugs will solve infinite life crises

immortality drugs will allow you to have an infinite life crisis

@craz_tom @GENIC0N Exuberance in the face of atrocity is a debt to reality and to God, a ledger that will be paid, ultimatley, during the last judgement, during the final hour

@LoserRigel @GENIC0N RETVRN

Future historians will view our psychiatric industry the same way we view witch trials and phrenology

@danfaggella And I could say that feeling things are defined by their impetus towards good feelings, and craft an ideology around that. If you don't agree with this ideology, what issues would you point out with respect to it?

Game theory is downstream of conatus. Will to Power and Will to Life and Elan Vital are also downstream of conatus (Bergson / Nietzsche admit this). I’m saying ‘non dead things are defined by their impetus to persist’. Levin himself considers this self evident. If it’s ‘primitive’ so be it lol.

Them: "Any future intelligence should ensure that all sentient life forms get to determine their own futures." Me: "Life is a set of interlocking / inter-depending patterns and processes. How do we make all the lions and deer happy, and the wasps and caterpillars happy?" Them:

@danfaggella Here's opus 4.8 with my bolted-on epistemic optimizations. It also converged on instrumental convg lol claude.ai/share/135d9e40…

"what makes the process of life ontologically good as opposed to neutral or even evil?" danfaggella.com/nu <-- my arguments against negative utilitarianism danfaggella.com/potentia <-- my argument for why potentia should be expanded / is, generally, a net boon (esp if we can ensure it can have a more blissful trajectory) [see 2 reasons listed in image, it's from said article] I'm open to any cruxes / arguments against these points! My ideas have evolved over time and will continue to evolve but at present this is some of how I think about this stuff so LMK

@danfaggella I can say every neural architecture we observe has a goal— to maximize its hedonic valence, so the ultimate good is that we ought to tesselate the universe with hedonium. Evobio-level just-so story, overfitting.

@tenobrus if this is a property of language itself and not of, say, training data, perhaps it's a factor in the Industrial Revolution kicking off in England

huge win for sapir-whorf today

@danfaggella I meant like what the wasp who lays its eggs in the caterpillar wants is good vibes, not to lay its eggs in the caterpillar. We can probably do better than the clever fuckape in terms of valence too! However, is hedonium tiling a failure mode or an ideal?

@LoserRigel I WHOLLY disagree that the generating function is superfluous without the underlying, rolling, self-overcoming process-of-life, no conscious stuff exists that which UNDERGIRDS is more important ultimatley, say I danfaggella.com/undergirding

@nelox @alex_verem its hilarious Grok dechudified so easily. I guess reality does have a liberal bias

@alex_verem Substitute lie for scam; this is what Grok answered:

The biggest scam humanity has collectively accepted as normal, according to Claude