Joscha Bach: meaning is instrumental to life
It's not that there is a meaning of life, but that meaning is instrumental to life.
Sarah Waker & Lee Cronin on Assembly Theory (and a definition of life)
Assembly theory makes the seemingly uncontroversial assumption that complex objects arise from combining many simpler objects. The theory says it’s possible to objectively measure an object’s complexity by considering how it got made. That’s done by calculating the minimum number of steps needed to make the object from its ingredients, which is quantified as the assembly index (AI).
[...]
[Lee Cronin and Sarah Walker] proposed a very general way to identify molecules made by living systems — even those using unfamiliar chemistries. Their method, they said, simply assumes that alien life forms will produce molecules with a chemical complexity similar to that of life on Earth.
https://www.quantamagazine.org/a-new-theory-for-the-assembly-of-life-in-the-universe-20230504/
Complex (technological) objects do not just appear spontaneously in the universe, despite popular folklore to the contrary. Cells, dogs, trees, computers, you and I all require evolution and selection along a lineage to generate the information necessary to exist.
[...]
Attempts to define life have so far failed because they focus on containing the concept of life in terms of individuals rather than evolutionary lineages.
[...]
We need to get past our binary categorization of all things as either “life” or “not.” We should not exclude examples based on naive assumptions about what life is before we develop an understanding of the deeper structure underlying the phenomena we colloquially call “life.”
[...]
Biological beings alive today are part of a lineage of information that can be traced backward in time through genomes to the earliest life. But evolution produced information that is not just genomic. Evolution produced everything around us, including things not traditionally considered “life.” Human technology would not exist without humans, so it is therefore part of the same ancient lineage of information that emerged with the origin of life.
Technology, like biology, does not exist in the absence of evolution. Technology is not artificially replacing life — it is life.
It is important to separate what is meant by “life” here as distinct from “alive.” By “life,” I mean all objects that can only be produced in our universe through a process of evolution and selection. Being “alive,” by contrast, is the active implementation of the dynamics of evolution and selection. Some objects — like a dead cat — are representative of “life” (because they only emerge in the universe through evolution) but not themselves “alive.
[...]
The canonical definition of technology is the application of scientific knowledge for practical use. [...] Technology relies on scientific knowledge, but scientific knowledge is itself information that emerged in our biosphere. It enables things to be possible that would not be without it.
Consider satellites. Launching them into space would not have been possible on our planet without Newton’s invention of the laws of gravitation. Newton himself could not have invented those laws if, centuries earlier, humanity had not come to understand the mathematics of geometry or constructed timekeeping devices that allowed us to track seconds. And of course, none of this could have happened if our biosphere had not evolved organisms capable of making abstractions like these in the first place.
Once the knowledge of laws of gravitation became encoded in our biosphere, new technologies were made possible, including satellites. Satellites are not launched from dead worlds or worlds with only microbial life. They require a longer evolutionary trajectory of information acquisition. You can trace that lineage within the history of our species, but arguably it should be traced all the way back to the origin of life on Earth.
Technology, in the broadest sense, is the application of knowledge (information selected over time) that allows things to be possible that are not possible in the absence of that knowledge. In effect, technologies emerge from what has been selected to exist. They are also what selects among possible futures — and builds them.
[...]
We are accustomed to thinking about technology as uniquely human, but in this broader definition, there are many examples across the biological realm. Just like the objects of life might include pencils and satellites, so too technology might include wings and DNA translation.
[...]
People might want to differentiate between biological evolution and the intentionality of humans when we build technologies. After all, software developers and companies choose to produce technology in a different way than ravens evolved wings to fly. But both fundamentally rely on the same principles of selection.
Arguably, the kind of selection humans do is much more efficient than natural selection on biological populations. It is more directed, which is only possible because we ourselves already are structures built across billions of years. We are bundles of possibilities refined by evolution and embodying the history of how we came to exist. The physics governing how we select what we create may be no different (other than by degree of directedness) than how we were selected by evolution. We are, after all, a manifestation of the very physics that allowed us to come to be.
[...]
The technologies we are and that we produce are part of the same ancient strand of information propagating through and structuring matter on our planet. This structure of information across time emerged with the origin of life on Earth. We are lineages, not individuals.
[...]
A natural extension of this evolutionary history is to recognize how “thinking” technologies may represent the next major transition in the planetary evolution of life on Earth.
[...]
What is emerging now on Earth is planetary-scale, multisocietal life with a new brain-like functionality capable of integrating many of the technologies we have been constructing as a species over millennia. It is hard for us to see this because it is ahead of us in evolutionary time, not behind us, and therefore is a structure much larger in time than we are. Furthermore, it is hard to see because we are accustomed to viewing life on the scale of a human lifespan, not in terms of the trajectory of a planet.
Life on this planet is very deeply embedded in time, and we as individuals are temporary instances of bundles of informational lineages. We are deeply human (going back 3.8 billion years to get here), and this is a critically important moment in the history of our planet but it is not the pinnacle of evolution. What our planet can generate may just be getting started. In all likelihood, we are already a few rungs down in the hierarchy of informational systems that might be considered “alive” on this planet right now.
We are 3.8-billion-year-old lineages of information structuring matter on our planet.
The Sun is massive
Every now and then I dip into cosmology 101, trying to improve my picture of—and my feel for—the Universe was find ourselves in.
Today I learnt that every second, the Sun burns off 4 million tonnes of mass—roughly 2 million cars.
It will do this for most of its 10 billion year “main phase”. During those 10 billion years, it’s total mass will decrease by only 0.07%. The Sun is massive.
And by the way: there are 250 million stars in the Milky Way galaxy. And there are hundreds of millions of galaxies in the Observable Universe. And the Universe itself may be infinitely large, whatever that means.
Awareness of things like this makes me more open to “wild” claims about the future of things on Earth.
Robert Skidelsky on aristocratic self-confidence
What distinguishes the aristocracy and gentry from the bourgeoisie in this new world of Evangelical piety and middle-class utilitarianism is nothing more than a nuance, a subterranean tradition hard to pin down. It is perhaps best expressed in the phrase ‘ high spirits: the lingering tradition of the dandy, of ‘sowing one’s wild oats’, of practical joking, of conspicuous display, of eccentricity, of idiosyncratic speech; all perhaps products of the self-confidence alone possessed by those born at the apex of an unchallenged social pyramid, who instinctively feel that anything they do must be justified.
Eric Bentley’s comment on Nietzsche’s Zarathustra is apposite: “The old aristocratic modes of feeling are to be grafted onto the thought of a new evolutionary futurism.”
G.A. Cohen on the conservative bias to retain what is of valuable, even in the face of replacing it by something of greater value
Conservative conviction, as I understand it, thereby exhibits a bias in favor of retaining what is of value, even in the face of replacing it by something of greater value (though not, therefore, in the face of replacing it by something of greater value no matter how much greater its value would be).
But there is a third idea, beyond preservation of the (intrinsically) valuable and the (personally) valued, in conservatism that warrants notice here, namely, the idea that some things must be accepted as given, that not everything can, or should, be shaped to our aims and requirements; the attitude that goes with seeking to shape everything to our requirements both violates intrinsic value and contradicts our own spiritual requirements.
Given things that are of value ought to be retained, but, beyond that, we need some things to be given quite apart from whether they have value other than because they are given. It is essential that some things should be taken as given: the attitude of universal mastery over everything is repugnant, and, at the limit, insane.
—Cohen (2012), Finding Oneself in the Other, "Chapter 8. Rescuing Conservatism: A Defense of Existing Value"
Hilary Greaves on definitions of existential risk
Generally, where “humanity” appears in a definition of existential catastrophe, it is to be read as an abbreviation for “Earth-originating intelligent sentient life”.
[...]
A little more fundamentally, the desideratum is that there exist, throughout the long future, large numbers of high-welfare entities. One way this could come about is if a takeover species is both highly intelligent and highly sentient, and spreads itself throughout the universe (with high-welfare conditions). But other ways are also possible. For example, takeover by non-sentient AI together with extinction of humanity need not constitute an existential catastrophe if the non-sentient AI itself goes on to create large numbers of high-welfare entities of some other kind.
Matt Ridley on innovation vs invention
Invention is the coming up with a prototype of a new device or a new social practice innovation. Innovation is the business of turning a new device into something practical, affordable and reliable that people will want to use and acquire. It’s the process of driving down the price; it’s the process of driving up the reliability and the efficiency of the device; and it’s the process of persuading other people to adopt it, too.
September assorted links
Understanding LLMs:
- Large language models, explained with a minimum of math and jargon
- How to make a simple GPT from scratch
AI policy:
- Samuel Hammond: Will there be major AI legislation? and AI & Leviathon
- Politico: How Rishi Sunak convinced the world to worry about AI and How Silicon Valley doomers are shaping Rishi Sunak’s AI plans
- Richard Danzig: Technology Roulette (2018)
AI alignment:
- Paul Christiano: Decoupling deliberation from competition. I think there's something off here; I'll write soon.
Philosophy:
- Wei Dei: Meta questions about metaphilosophy
- Paul Grice: The cooperative principle
Misc:
- Max Roser: How many animals get slaughtered every day?
Jerry Z Muller on left and right-wing totalitarianism in the early 20th century
The choice of the proletariat rather than the Volk or nation reflected a deeper divergence between intellectuals of the left and right, a divergence whose origin lay in the eighteenth century. It hinged upon differing answers to the question "What is the source of the purposes that men ought to hold in common, and of the institutions that embody those purposes?"
Those who believed that the ultimate source of such purposes lay in "reason," which was capable in principle of providing answers for men and women in all times and places, arrayed themselves as the party of Enlightenment. Despite great internal divergences, it was the universalist, cosmopolitan thrust of its thought that united what Peter Gay has called "the party of humanity"-a designation that recaptures the self-understanding of the philosophes. Though it may have shifted the locus of reason in the direction of the methods of natural science, the Enlightenment maintained the far older belief in a rational universe with a necessary harmony of values accessible to human reason. The theory and practice of enlightened absolutism, revolutionary republicanism, and the application of the Code Napoleon to non-French nations within the Napoleonic empire all shared the premise that because men were fundamentally the same everywhere, there were universal goals and universalizable institutions for their pursuit. In its universalism--its commitment to a proletariat that was to know no fatherland and whose interests were held to be identical with those of humanity--Marxism and its totalitarian communist variant were intellectual descendants of the Enlightenment. Indeed, it was the premise and promise of universalism that made communism disproportionately attractive to intellectuals who sprang from ethnic and religious minorities.
The intellectuals who placed their hopes on totalitarian movements of the right usually descended from what Isaiah Berlin has termed the "Counter-Enlightenment." Its advocates formed no party; what they shared was a skepticism toward "the central dogma of the Enlightenment," the belief that the ultimate ends of all men at all times were identical and could be apprehended by universal reason. The orientation of the thinkers of the Counter-Enlightenment was usually historicist and particularist. They regarded the attempt to discover universal standards of conduct as epistemologically flawed, and the attempt to impose such standards as a threat to the particular historical cultures from which individuals derived a sense of purpose and society a sense of cohesion. The thinkers of the Counter-Enlightenment regarded the variety of existing historical cultures as both inescapable and intrinsically valuable and suggested that the multiplicity of historical cultures embodied values that were incommensurable or equally valid. For thinkers of the German Counter-Enlightenment such as Johann Gottfried von Herder and Justus Möser, cosmopolitanism was the shedding of all that makes one most human, most oneself. Although by no means uniform in their politics, they all resisted the attempt at a rational reorganization of society based upon purportedly universal and rationalist ideals
In countering the claims of the Enlightenment, the most original thinkers of the Counter-Enlightenment enunciated claims that represented a major departure from the central stream of the Western tradition, which had held that problems of value were in principle soluble and soluble with finality. The Counter-Enlightenment, by contrast, maintained that the traditions which gave a group its identity and which were expressed in its culture were not themselves rationally grounded and could not always be rationally justified. The value of a culture or institution was conceived of as deriving from its history, from its role in the development of a particular group. Proponents of the Counter-Enlightenment such as Möser or Burke revived the argument of the Sophists that moral order was a product of convention that varied from group to group. To this they added the perception (pioneered by David Hume) that emotional attachment to an institution—what Burke called "reverence"—was often a product of the institution's longevity. This social-psychological perception lay at the heart of the argument for the functional value of continuity, a theme that provided the basis for numerous variations among later generations of conservative thinkers.
—The Other God That Failed (1988)
Richard Danzig on technology roulette and the U.S. military
This report recognizes the imperatives that inspire the U.S. military’s pursuit of technological superiority over all potential adversaries. These pages emphasize, however, that superiority is not synonymous with security. Experience with nuclear weapons, aviation, and digital information systems should inform discussion about current efforts to control artificial intelligence (AI), synthetic biology, and autonomous systems. In this light, the most reasonable expectation is that the introduction of complex, opaque, novel, and interactive technologies will produce accidents, emergent effects, and sabotage. In sum, on a number of occasions and in a number of ways, the American national security establishment will lose control of what it creates.
A strong justification for our pursuit of technological superiority is that this superiority will enhance our deterrent power. But deterrence is a strategy for reducing attacks, not accidents; it discourages malevolence, not inadvertence. In fact, technological proliferation almost invariably closely follows technological innovation. Our risks from resulting growth in the number and complexity of interactions are amplified by the fact that proliferation places great destructive power in the hands of others whose safety priorities and standards are likely to be less ambitious and less well funded than ours.
Accordingly, progress toward our primary goal, superiority, should be expected to increase rather than reduce collateral risks of loss of control. This report contends that, unfortunately, we cannot reliably estimate the resulting risks. Worse, there are no apparent paths for eliminating them or even keeping them from increasing. The benefit of an often referenced recourse, keeping “humans in the loop” of operations involving new technologies, appears on inspection to be of little and declining benefit.
https://www.cnas.org/publications/reports/technology-roulette
Alien invaders, or children?
If we think of digital minds as alien invaders, we fight them to our last breath 1.
If we think of digital minds as our children, we would raise them with care and wish them well. And we could expect them to wish us well too.
Which is the better frame?
I think it mainly depends on: (a) How much they share our values. (b) Whether they are capable of living great, flourishing lives. Are they conscious?
Those who prefer the alien frame may also think that biological relation matters. Why? (i) Loyalty to one's species. (ii) Conservatism: ways of life are good just because they exist; therefore its bad if they are lost, and their replacement by better things doesn't automatically make up for that.
What are the other cruxes?
-
A consequentialist who holds an impartial theory of value might not. They might think that letting the aliens win would create a more valuable future. ↩
Reading suggestions
Someone asked for a book suggestion. My reply...
It's hard to think about anything other than AI these days. The books are mostly out of date.
For AI, I suggest:
- Samuel Hammond's blog
- LLM explainer
- Sparks of Intelligence
- We can't compete
- A rehash of Bostrom's Vulnerable World paper, notably written by an influential SV figure with background in the Open Science movement. He's late to the party, but at least he's here now.
Things on my mind:
- How do we regulate and police our way to a better offence/defence balance, without wrecking civil liberties? We can surely do better than totalitarianism.
- Will the policy response in 2023/24/25 slow things down for several decades? Hard to tell.
- Should we think of digital minds as alien invaders, or our children, or something else? cede the future to as we do to children, rather than than see as alien invaders.
- Very new politics coming soon. How to prepare?
Big news of the year: most world leaders basically get it now. This has happened faster than I expected.
In the "not-AI" section, I liked Professor of Apocalypse and The Other God That Failed, both by Jerry Z Muller.
More and more, I read with GPT-4 open, asking questions as I read, as I might with a private tutor. Pay for ChatGPT if you didn't already.
Robin Hanson on the future as reality (read by my digital avatar)
The audio and video is generated from a text input, by Heygen.
Their system was trained with 30 seconds of video from my webcam.
Here's Nietzsche on philosophers:
ChatGPT is a junior engineer today; a senior designer and engineer next year
If you use several Google accounts, Google Meet does not automatically switch to whatever account has permission to join the call. It's annoying.
Today I hired a team and made a browser extension to solve this.
It took us less than 30 minutes to get a working prototype. It took a further 3-4 hours to make a "production-ready" version for release on the Chrome Web Store.
I served as UX designer, engineering manager and head of QA. GPT-4 was our lead developer, and wrote roughly all of the code, to a higher standard than I would have 1.
How will GPT-5 help us make faster progress?
Here's my guess: (1) The workflow will be: I give instruction, it updates codebase, it checks the result in browser, fixes codebase if necessary, then asks me to review. 2 (2) It will require less of my "expert direction" on how to approach some of the engineering tasks. (3) It will reply as quickly as GPT-3.5. (4) It will help me create the logo, screenshot and demo video for the Chrome Web Store. (5) It will help me design the user interface. (6) It will proactively suggest feature ideas and possible UX issues.
I expect most of these improvements will be available by the end of 2024. With them, I'd be able to create this extension in less than an hour.
Today, my background in software design and engineering was a necessary condition to deliver a great result within a couple of hours (see the transcript). How much of this expertise will be needed in 2024? 2025?
Prediction: by December 2025, my wife (who has no software design or engineering experience) will be able to create the same extension in less than a day.
-
The full codebase is some 500 lines. Less than 5 of these were entirely written by me. Perhaps 30-50 were started by me, and completed by our intern (Github Copilot). ↩
-
Open Interpreter (released last week) now offers this workflow for writing blocks of code, but until these systems can "see", I'll need to do a bunch of the software testing in browser. Multimodal LLMs are due before the end of this year. ↩
Peter Singer on the non-natural origin of the axiom of impartial benevolence
According to Singer, to the extent that this intuition is not adaptive, we can think of it as "a truth of reason" rather than another value that emerges from competition and selection.
If we apply this to the dualism of practical reason, then what we have is, on the one hand, a response—the axiom of universal benevolence—a response that clearly would not be likely to have been selected by evolution, because to help unrelated strangers—even at some disadvantage to yourself, where there's a greater benefit to the unrelated strangers—is not a trait that is likely to lead to your improved survival or the improved survival of your offspring. It's rather going to benefit these unrelated strangers who are therefore more likely to survive and whose offspring are more likely to survive. So that doesn't seem like it would have been selected for by evolution, which suggests that maybe it is a judgement of our reasoning capacities, in some way; we are seeing something through reason.
Now, if we compare that with the egoistic judgement that I have special reasons to prefer my own interests to those of strangers, it's more plausible to think that that would have been selected by evolution. Because, after all, that does give you preference for yourself and your offspring, if you love your offspring and care for them.
So if we have these two conflicting judgments, then maybe we can choose which one by saying: just as in the case of adult sibling incest, we debunk the intuition by saying, “Well, that's just something that evolved in our past and that doesn't really give us reasons for thinking the same thing today,” maybe we can say that also about the intuition behind egoism, but not about the intuition behind universal benevolence, which therefore gives us a reason (not a conclusive or overriding reason) for thinking that it's the axiom of universal benevolence that is the one that is most supported by reason.
Nope. The way we reason is a product of natural selection. You don't get to pick some bits you like and say that these are universal truths.
AI will improve a lot, very soon
It is hard to feel "in my bones" just how much is now "baked in" and coming this year, next year, the year after. I am starting to write down concrete predictions, as part of the AI OODA loops I'm running this autumn.
Meantime: a few things have helped me "feel it" in the last week or so: v0.dev, Open Interpreter, and this note from Michael Nielsen:
As an aside on the short term – the next few years – I expect we're going to see rapidly improving multi-modal foundation models which mix language, mathematics, images, video, sound, action in the world, as well as many specialized sources of data, things like genetic data about viruses and proteins, data from particle physics, sensor data from vehicles, from the oceans, and so on.
Such models will "know" a tremendous amount about many different aspects of the world, and will also have a raw substrate for abstract reasoning – things like language and mathematics; they will get at least some transfer between these domains, and will be far, far more powerful than systems like GPT-4.
This does not mean they will yet be true AGI or ASI! Other ideas will almost certainly be required; it's possible those ideas are, however, already extant. No matter what, I expect such models will be increasingly powerful as aids to the discovery of powerful new technologies.
Regulation is likely to be a significant headwind, but a lot more "transformatively useful" stuff is landing soon regardless.
My software development workflow will speed up a lot, again, within the tools that will drop within a year.
Peter Thiel on the Hegelian tension of AI
It’s deranged decentralisation vs totalitarian centralisation. We need to find a third way.
Peter Thiel on the Darwinian view of alignment
The core axiom of the Darwinian view of the world: there is no such thing as a selfless being. Therefore the alignment problem is fundamentally difficult.
Asmir Gračanin et al. on the function of tears
In the inter-personal domain, tears can be considered as a signal that conveys information about the helplessness of the crying individual resulting in increased motivation of observers to react with pro-social behaviour. However, the inter-personal effects of crying are also not always consistent. Although evidence suggests that the perception of tears generally results in helping behaviours, strengthening of social bonds and a reduction of aggression, there is convincing anecdotal evidence that (particularly acoustical) crying may also sometimes evoke irritation and even aggression and violence. The precise determinants of the reactions of others to crying still wait to be disclosed. Nevertheless, it seems obvious that the personality of the observer, the specific antecedent and the perceived appropriateness of crying, as well as well as the relationship between crier and observer may all play a role.
https://www.tandfonline.com/doi/full/10.1080/02699931.2016.1151402