Overall I like this take, and I feel like it aligns closely with my own beliefs about the future of technological progress. With that said, I feel like this and

Overall I like this take, and I feel like it aligns closely with my own beliefs about the future of technological progress. With that said, I feel like this and most discussions about technology are far too limited in what they consider to be "technological". To paraphrase Ursula Franklin: technology is systems, not just the artifacts that we typically consider "technology". As an example, consider "the car". There is the artifact of an automobile, for sure, but when we say "the technology of cars" we are referring to something which is a superset of the collection of physical automobiles: we are also tacitly including all of the human protocols which have been put in place to enable the automobile artifacts to "function properly" (move a human body from place A to place B safely and quickly). Road design and maintenance, drivers ed, traffic engineering, urban planning etc. are all components in the larger technological system that defines the capabilities of "the car". I'm not saying that it's wrong to refer to subsystems of this technology as "technologies" of their own, just that it's a more difficult taxonomical problem than many are willing to acknowledge. There are no artifacts functioning in a vacuum; all of our tools are embedded in and partially defined by the human systems around them. I say this because I think it changes what it means to be a "techno-optimist". To be optimistic about the capabilities of technology to produce beneficial outcomes is more than just being optimistic about the capabilities of the artifacts ("we will develop more effective vaccines and more efficient batteries"), but also being optimistic about human systems. I do think that Vitalik has this somewhat in mind in this post; his discussion of differential development, as well as the role of governance systems gets towards this. However I wish he was more explicit about it because "technological progress" takes on a radically different flavor from this perspective. To some extent it becomes tautological: progress in human techno-systems is progress in human society more generally. However it also lays out additional axes to consider development along. For example: moving towards his solarpunk utopia north star could conceivably occur with no or little change in the current capabilities of our tooling, and instead result from a radical reorganization of human society. Extrapolating further you might conclude that a true d/acc techno-optimist should spend less time developing cryptographic consensus mechanisms, and more time radicalizing the folks at your community garden to become self-sustaining and mutually supportive. This is extreme, and probably the best way forward is "both/and", but I think that it gets underplayed in any of these discussions about "technology", resulting in a somewhat limited scope of action items. When it comes to AGI I think this perspective is even more salient, because it unbundles the single "AGI" entity into a collection of subsystems and allows us to more effectively game out what the probable AGI futures are. Indeed, my own somewhat contrarian take is that AGIs already exist in the form of the human organizations (states, corporations, etc). These are decision-making entities composed of humans, but I think it's not unfair to consider them as a type of hive-mind entity which deserves status as "intelligent agent". The popular AGI anxiety, that we will create a novel type of agent which will "function" free of human systems entirely, seems unlikely to me: given how little we know about what a "mind" is, it seems like a stretch to assert that one could be engineered in silicon so easily. Rather I think the more probable outcome is Vitalik's AI-supercharged totalitarian state: instead of carving itself free of our systems I think it's likely that an old version of an AGI (the state or corporate apparatus) upgrades itself using the new computing paradigms available to it. This has happened many times before: the invention of statistics and the big data revolution(s) show us what happens when computing and decision making infrastructure embeds itself in the governing technology. Looking back to these previous "AGI apocalypses" we can more effectively extrapolate and plan for the future. Rather than being an entirely new problem space (Vitalik compares it to the origin of life; to me this is an enormous stretch), it becomes one with which we are exceedingly familiar.