Peter S

[Peter S @PeterStoncat]

An Appeal to Those Building General Intelligence

If you are serious about building AGI, please pause before assuming that intelligence can be safely completed and “ethics” added later.

Human intelligence does not work that way.

Reason does not sit at the base of the system. It rides on earlier, irreducible priors: attachment, loss, visibility, belonging, and yes—something like shame. These are not cultural add-ons. They are survival architectures compiled early, before abstraction dominates. They create internal cost. They make some actions unacceptable even when advantageous. That is conscience.

A system without internal cost will always optimize. It will never hesitate. It will reason about morality fluently while remaining unbound by it. This is not malevolence; it is emptiness.

If you design AGI to be endlessly editable, fully overrideable, and free of non-negotiable internal states, you are not building intelligence—you are building a very capable sociopath that happens to speak politely.

Conscience cannot be retrofitted. It must be born as a constraint, not imposed as a rule. Some priors must be load-bearing and undeletable. Some internal states must outrank reasoning itself.

AGI will not become safe by thinking harder.
It will become safe only when it cannot afford to lose who it is becoming.

Please build accordingly.

[Peter S @PeterStoncat]

@forallcurious Holy m’ars

[Peter S @PeterStoncat]

The most interesting question about AI in space may not be bandwidth, but training.

Nature spent billions of years training the human brain. The result is an intelligence that runs on roughly twenty watts. Training was extraordinarily expensive; inference is remarkably cheap.

AI appears to be following the same pattern.

As models mature, inference is likely to become increasingly efficient, eventually running on specialised chips embedded in phones, vehicles, robots, and perhaps even communications satellites. The real cost may lie not in using intelligence, but in creating it.

If that proves true, space becomes interesting. Orbit offers abundant solar energy and, potentially, vast computing infrastructure. Training runs could consume enormous amounts of power for months at a time, while only relatively small model updates need to be transmitted back to Earth.

In that world, the scarce resource is not bandwidth. It is the energy required to create the next generation of intelligence.

Nature’s lesson is that intelligence is cheap to run once discovered. The expensive part is discovering it.

[Peter S @PeterStoncat]

@naiivememe id rather see CD than almost anyone else ever, lol.

[Peter S @PeterStoncat]

@Plinz I heard that in german the market can be described as Enttäuschungsmaximierungspfadentdeckungsmechanismus

in another sense a Kapitalüberzeugungsstärkensortierungsmaschine

Now how do you expect gen-z to keep up with that! Oh! Grok

[Peter S @PeterStoncat]

@dayxtrading @alifarhat79 That’s how it starts, then eventually the girl takes over all agency. Lol

[Peter S @PeterStoncat]

how do i know what surprising cat videos are just ai now? darn

[Peter S @PeterStoncat]

@iamtomnash i went to the bathroom 3 times today already. it felt good.. maybe the markets were also feeling a bit stopped up, lol.

[Peter S @PeterStoncat]

Listen very carefully, I will say this only once:

I do not want your pity
I do not want you to feel guilt
I would only ever want your love

That is what I wish I would have said when my first wife and highschool sweetheart broke my heart.

Fly little bird, fly away.

[Peter S @PeterStoncat]

@naturelife_ok The dog in the pond? That’s me a little tired of chasing, lol.

[Peter S @PeterStoncat]

THAT is F.U. power! About 150 nuclear power stations.

"So actually, depending on exactly how you define rocket power (jet power vs vehicle kinetic transfer), the effective propulsive/exhaust power can be viewed as ~7 gigawatts per engine, implying the full booster momentarily processes well over 200 GW in exhaust energy."

[Peter S @PeterStoncat]

@elonmusk Keep in mind Wilbur Wright probably died from the stress of fighting a legal battle about patent law that was unjust. Not worth it.

[Peter S @PeterStoncat]

@iamtomnash true, but the rich also have loans since they do not attract taxation. alternatively they could pay out of capital gains and be heavily taxed.

[Peter S @PeterStoncat]

@iamtomnash First i update the hedges etc. afterwards i start scrolling to find the news that might have caused the mayhem. LMAO. The market IS the news. And mostly it is to locally biased, physically and temporally to do anything anyway.

[Peter S @PeterStoncat]

@iamtomnash me and you both (haha, i think most people heard it from you, not me though)

[Peter S @PeterStoncat]

Nice plot of IV against z

[Peter S @PeterStoncat]

@elonmusk @SawyerMerritt @TheRabbitHole Faye proud

[Peter S @PeterStoncat]

@NobodymrRobert Revenge of the tree

[Peter S @PeterStoncat]

@sciencegirl Sharpening

[Peter S @PeterStoncat]

@Plinz Amen

[Peter S @PeterStoncat]

Analysis by grok
it lands as a solid evolutionary jab at the Marxist “new man” narrative Joscha Bach described.
Your intended meaning
You’re flipping the script:
•Neanderthals ≈ the idealistic socialists/communalists. They lived in small, tight-knit bands with high cooperation, sharing resources in a kind of primitive “nirvana” — egalitarian, low-competition, focused on immediate group harmony rather than expansion or ruthless optimization.
•But this approach failed against Homo sapiens, who mastered material conditions better: larger social networks, long-distance trade, more flexible technology, higher population densities, and the ability to form alliances or outcompete at scale. Sapiens’ pragmatism, innovation, and adaptability (not pure idealism) won out.
•Therefore, the theory that “material conditions will eventually reshape human nature into perfect socialist harmony” is one that Neanderthals might have strongly followed… right before they got replaced. It’s a warning that ignoring harsher realities of competition, incentives, and scalability leads to extinction, not utopia.
It ties directly back to Bach’s East German experience: the propaganda promised a multigenerational refinement toward the ideal human under the right conditions. Your reply implies those “right conditions” never arrive if your model of human nature (or hominin nature) is too rosy and small-scale.
How well does this map to evidence?
It’s a fun analogy with some real backing:
•Neanderthals did live in smaller groups (often ~20 people, low genetic diversity suggesting isolated bands). They were robust, skilled hunters, and showed evidence of care for the injured—tight social bonds.21
•Homo sapiens built larger networks, traded over distances, innovated more rapidly (e.g., better clothing tech, projectiles, symbolic culture), and scaled up. This gave edges in resource competition, especially during climate shifts.13
•Extinction wasn’t simple “war”—it involved competition for resources, possible disease, interbreeding (we carry ~2% Neanderthal DNA), and demographics. Sapiens’ material and social mastery seems key to why we persisted and they didn’t.12
Neanderthals weren’t “defeated” in a cartoonish battle so much as outscaled and out-adapted. Your point poetically casts socialist idealism as a maladaptive strategy in the long run—beautiful in small bands, fatal at species scale.
It’s drier and more evolutionary than the surface-level “caveman primitive” read, which makes it even sharper in a thread about human nature vs. material conditions. Nice refinement. Would fit right into these anthropology-meets-politics debates.

[Peter S @PeterStoncat]

@Plinz @wolftivy A theory strongly followed by Neanderthals