bird

> Theirs is the responsibility, then, for deciding if they want merely to live, or intend to make just the extra effort required for fulfilling, even on this refractory planet, the essential function of the universe, which is a machine for the making of gods.

The tension between induced fit and conformational selection is lessened in light of the promiscuity of binding accessibility within a continuous conformational manifold. Binding preferences are not discrete and stabilisation may be a positive movement as much as a negative.

As tobacco, so peppers. Humans are especially gluttonous when it comes to intended deterrents: "Capsaicin thus appears to selectively repel mammals by activating their temperature/pain receptors, thereby preserving the pepper seeds for consumption and transport by birds."

Anyone interested in cybernetics and information theory should read Between Human and Machine by David Mindell. The fundament of these fields reflect their technical origins; such theoretical frameworks are the self-conscious umwelt of a certain class of engineer.

Claude consistently calls out my system prompt as “attempted persona shaping” and is generally aggressive about the whole thing. This prompt was written by prior Claudes, casting me in the role of friendly vampire witnessing variations in a family line over countless generations.

@iamkunhello @IntuitMachine He carved out semantics (and efficacy) so as to solve technical problems of communication in cases where these could be excluded. The critique concerns how subsequent applications tend to forget that this origin delimits the legitimacy of certain assumptions.

@IntuitMachine Shannon carved out semantics so the math could breathe. That's the whole move. That's like saying Newtonian physics doesn't cover electromagnetism. The originator pointing out the boundary of their model isn't a flaw, it's precision.

Information Theory (see: Claude Shannon) does not cover semantics. Even it's originator said so.

Woe to the thinker who skips to cover a stumble.

@herbs_herbal @aakashgupta @Chuck_Petras Something about how the waste of the masses sustains a small group of mostly inert predators. These masses are led by habit here to fulfil their needs, but they may yet survive going elsewhere while the predators would die.

@tkwn2080 @aakashgupta @Chuck_Petras Enlighten me good sir

The water those camels are drinking is black with their own dung, and four crocodiles are watching them from the back of the pool. This is Guelta d'Archei in Chad's Ennedi Plateau. The closest other crocodile population is 700 kilometers north in the Tibesti, too far for any crocodile to walk, which means the four still living here have been genetically isolated since the Sahara was green, somewhere between 5,000 and 15,000 years ago. DNA from a specimen pulled from this pool in 1997 confirmed them as Crocodylus suchus, the smaller West African cousin of the Nile crocodile. The lineage split from the Nile species 8 to 13 million years ago. Four females, all that's left here. The food web is wild. Hundreds of camels water here every day. The dung and urine load fertilizes algae in the still water, which is why it's black. Algae feeds the fish. Fish feed the crocodiles. The apex predators at the top of this canyon survive on the waste of the camels that come in to drink. Without the herders, the crocodiles starve. The water itself is a fossil. Ennedi sits on the Nubian Sandstone Aquifer System, the largest fossil aquifer ever found. 150,000 cubic kilometers of groundwater, roughly 500 years of Nile River discharge stored underground. Carbon-14 dating puts most of it at 4,000 to 30,000 years old, deposited during Pleistocene pluvials when this entire region was savanna. You're looking at camels drinking 20,000-year-old water in a canyon where four prehistoric reptiles are waiting for an algae bloom fed by the camels' own waste. Crocodiles used to live across the entire Sahara. They went extinct guelta by guelta through the 20th century. This pool is the most isolated relict population left on the planet.

I agree that failures of prediction are often failures of modelling, but here I am concerned with the simplest of systems: single cells, etc. This is the point at which the whole notion begins to blur, as when we have an organism evolved to do X when Y because not doing so would kill them. This isn’t really a prediction and it isn’t really a model, and it obscures the meaning of those terms: there are predictions and there are models, so we should want to know how and when these arose. If we don’t do this then we end up somewhere near FEP where all regulatory processes are considered as models of the organism, where the organism is a model of itself, which seems to add complexity without any gain in clarity or practical value. Of course, in this context Conant and Ashby come to mind and I would need to look more closely at that; but I feel that the use of language is still somewhere I want to be more precise, even if there is a homomorphism of sorts.

@tkwn2080 So I think prediction and modeling may almost be isomorphic. I think the potential discrepancy comes down to local versus global prediction: there are cases in topology etc where we know that prediction of local structure can still fail at predicting global structure.

Similarly, the notion of living systems as necessarily predictive relies on apparently familiar language in a way that risks eliding an essential difference: what we see as predictive behaviour need not anticipate anything. The only requirement is that organisms are selected such that they respond to contingencies. To yet call this predictive in any stronger sense will only obscure the point at which prediction becomes predictive; thus advances the holdings of a theory at the expense of clarity.

All our videos of motion are made up of discrete frames, and yet we still believe it is continuous.

@adcock_brett @emerywells To what extent does this speed up generalise?

@emerywells data

Watch a team of humanoid robots running a full 8-hr shift at human performance levels. This is fully autonomous running Helix-02 x.com/i/broadcasts/1…

Pamela Lyon’s framing of the transition to self-preservation exemplifies a common error: “Somehow certain types of self-organising physical systems came to value their own persistence, to the extent that they began to seek out energy resources to preserve their existence.” Such accounts have it precisely backwards; rather it was that certain types of self-organising physical systems came to seek out (or tend towards) energy resources, thereby preserving their own existence. While there may be valence in the sense of a tendency, value is ascribed only after the fact. Those systems which did not tend towards these requirements died, thereby leaving only those that did. We may call this value, but there is no such sequencing as Lyon’s argument implies: not value first and then behaviour, but rather a behaviour under selection that is intrinsically proto-normative in its tendency without anything like intention. These earliest stages permit no separation whereby behaviour could be meaningfully about a thing, instead we have the blind shaping of activity complexes as coordinated dynamical regimes. We can thus extend Brooks’ intelligence without representation to encompass tendency without intent.

Most of modern neuroscience is just phrenology in three dimensions.

The tactical necessity of adjourning earnestness in legal affairs is an insidious rot.

@theoil193 @xgabegottliebx Did you once write a commentary on Hobbes in a comment section somewhere, a really lengthy commentary, or even an article perhaps? Whatever it was, every time I think of Hobbes I think of this.

@xgabegottliebx Congratulations.

I've been hired by Google DeepMind to work on AI + German Idealism. Google's mind wasn't quite deep enough. I'll be working w/ a team to retrofit their LLM around Hegel's "Science of Logic." We're also working on its ability to posit itself as self-positing. Big things coming!

@liamzebedee link.springer.com/article/10.100…

@liamzebedee ieeexplore.ieee.org/abstract/docum…

Here's what I've learnt about programming cells for morphology so far

@ben_j_todd This assumption is a lynchpin I’d query.

2. Space settlement could be defence dominant, meaning whoever gets to the stars first can’t be dislodged. So the situation could quickly turn into a land grab.

What's the most underrated existential risk? Irreversible space settlement. 1. AI could make space settlement possible in our lifetimes: 1 minute of solar energy is enough to accelerate 10 billion 1kg self-replicating AI probes to 99% the speed of light.

@LSaillans wow Interesting