But the truth is that no one has power.
I agree with what you're pointing at, but not with this statement. I think it's worse than this. There are (groups of) people who (collectively) know how to make (some) things better. But The Power is held collectively by those you label The Powerless, and they lack the skills or drive to even know how to choose the right priesthood-holders to trust. As a result we spend an awful lot of time and effort tying our hands and shooting our feet, while competing would-be priests shout ineffectually at one another, regardless of who has a record of having made correct predictions before. Our ancestors may not have understood the natural world, but if someone showed up and made the right guesses about the behavior of some eldritch god vastly more often than anyone else could do, they would have been either elevated to leadership or feared/hated/condemned for black magic. We lost that skill, I think, in favor of playing social status games, the moment it sunk in that the natural-world-threats felt distant.
Aka: In practice we are sex-obsessed murder-monkeys and all of this is way above our pay grade.
I had exactly one teacher, a professor in college, who understood and acted on this idea. His grading formula was such that acing the final always meant you got an A. But, the better you did on homework and tests before that ('achievement points'), the less the final exam counted for. And the more effort you put in (homework, office hours, class participation), the more leniently your exams would be graded ('effort points'). And the more effort the class as a whole put in, the more leniently everyone's exams would be graded. But I think something like that only works if you're willing to actually let students fail.
On grade level books and learning to read: Wow that's some serious insanity I had not been aware of. When I was in 6th grade one of the vice principles had a bookshelf in his office full of (I think donated?) books. Anyone could borrow one, and keep it if they read it and wrote a report on it. The first one I borrowed was the a compilation of Plato's dialogs and the Republic - that was my first real introduction to Philosophy as an academic subject. Also, my brother-in-law was reading chapter books at 3. And my whole first grade class was expected to be reading chapter books.
On even 'gifted' schools mostly not wanted kids to get far ahead: Many (in my limited personal experience, most) teachers are not experts in the subjects they're teaching. A 7th grade math teacher is, if you're lucky, an expert at teaching 7th grade math to typical 7th graders. They hopefully have a familiarity with 8th grade math and should be able to explain it, but I expect many would struggle to explain 10th grade math. Similarly, I wouldn't expect them to be prepared to teach 7th grade math to even very bright and curious 4th graders - I would think doing that may require a deeper level of mathematical understanding and also social/psychological awareness in order to adapt the approach to where the students are.
AI infrastructure seems really expensive. I need to actually do the math here (and I haven’t! hence this is uncertain) but do we really expect growth on trend given the cost of this buildout in both chips and energy? Can someone really careful please look at this?
This is not a really careful look, but: The world has managed extremely fast (well, trains and highways fast, not FOOM-fast) large-scale transformations of the planet before. Mostly this requires that 1) the cost is worth the benefit to those spending and 2) we get out of our own way and let it happen. I don't think money or fundamental feasibility will be the limiters here.
Also, consider that training is now, or is becoming, a minority of compute. More and more is going towards inference - aka that which generates revenue. If building inference compute is profitable and becoming more profitable, then it doesn't really matter how little of the value is captured by the likes of OpenAI. It's worth building, so it'll get built. And some of it will go towards training and research, in ever-increasing absolute amounts.
Even if many of the companies building data centers die out because of a slump of some kind, the data centers themselves, and the energy to power them, will still exist. Plausibly the second buyers then get the infrastructural benefits at a much lower price - kinda like the fiber optic buildout of the 1990s and early 2000s. AKA "AI slump wipes out the leaders" might mean "all of a sudden there's huge amounts of compute available at much lower cost."
1Figuring out what a startup should say to investors is strangely useful for figuring out what it should actually do. Most people treat these questions as separate, but ideally they converge. If you can cook up a plausible plan to become huge, you should go ahead and do it.
If you're not a software company, and what you want to do requires steel in the ground, then any workable plan to become huge will realistically require 3-4 years each in the lab, pilot, demo, and FOAK phases, largely in series, and will often benefit from the founders stepping down as CEO quite early in favor of someone with much more direct industry experience, and if you're honest about that many VCs will run away.
As in, because you might have to raise at a lower number in the future, you should raise at a lower number than that now, so you don’t have a ‘down round.’ Or because you couldn’t handle having the cash.
As you explain later, the first part of this would be nonsense if the second part weren't so important. AKA, if only the founders have the discipline to not increase spend rate beyond necessity and instead use the money to increase runway and still follow an optimal path to growth, instead of inefficiently chasing faster growth by spending more and just assuming more funding will be available when needed, this would not be such a problem.
Also it's not about having a down round, necessarily. It's sometimes about needing one at all. I've met people whose shareholders forced their companies to wind down instead of allowing a down round, even if the down round would likely have led to a successful exit later, because e.g. the shareholder was trying to raise their own next fund and a down round on their record would have made it harder.
Fair enough. If nothing else, it's best to state where you think the min value, max value, and midpoint are, and ideally put error bars around those.
Or, at least, you can state outright you think the midpoint and max are far enough away as to be irrelevant distractions to some particular practical purpose. Or that you expect other factors to intervene and change the trend long before the later parts of the sigmoid shape become relevant.
To add: It is in principle very easy for people to make equivalent prediction errors in either direction about when a particular exponential will level off, and to be wrong by many orders of magnitude. In practice, I usually encounter a vocal minority who happily ignores the fact that the sigmoid even exists, and a larger group who thinks that leveling off must be imminent and the trend can possibly continue much longer. The cynic in me thinks the former group tends to start getting believed just in time to be proven wrong, while the latter group misses out on a lot of opportunities but then helps ensure the leveling off has less catastrophic consequences when it happens.
I'm curious: was there a particular (set of) sigmoid(s) you had in mind when writing this post? And particular opinions about them you haven't seen reflected in discussions?
Most of the times I've used sigmoid in my own modeling/forecasting have been about adoption and penetration of new technologies. Often the width of the sigmoid (say the time to get from 1% to 99% of the way from min to max) is relatively easy to approximate, driven by forces like "how incumbent institutions are governed" (yes, this is critical even for most extremely disruptive innovations). The midpoint and maximum are much harder to anticipate.
This is absolutely true. However, actually using it effectively requires having a sufficiently good, principled reason for thinking the limit is in some particular place, or will be approached on some particular timeline. When I've looked at many (most?) real-world attempts to forecast when some particular exponential will start looking like an s-curve, they're usually really far off, sometimes in ways that 'should' be obvious, even if the forecasting exercise itself is instructive.
I feel like this underestimates the difference between what you're citing, valuation growth over 2 years post-YC, and the Garry Tan quote, which was about weekly growth during the 3 month YC program. I also wish the original Garry Tan claim were more specific about the metric being used for that weekly growth statistic. In principle, these aren't necessarily mutually exclusive claims. In practice, I'd expect there's some fudging going on.
I can imagine something like the following: Companies grow faster with less investment, reaching more revenue sooner because of GenAI. But, this also means the company has fewer defensible assets, and less of a lead over competitors, so the valuation is lower after 2 years. AKA potentially the cost of software innovation is going down, speed is going up, and there's less of an advantage to being first because it's getting easier to be a fast follower. In a world where it's possible-in-principle to think about one person unicorns, then why should software companies ever have high valuations at all, once enough people know what they're trying to build?
I'm curious what effects with would have, if true. If we end up in a place where an individual can build a $10-20M company, practically on their own, in months, with only a seed round, but can never get to $100M or $1B, how does this affect startup funding models? Pace of overall innovation? I could see this going really well (serial founders have time in their careers to build 50 companies, cost to access new software products drops) or really badly (VCs lose interest in software startups, innovation slows to a crawl), or any number of other ways.
Or, also not mutually exclusive with the above: Maybe GenAI-2023 is sufficiently different from GenAI-2025 that we shouldn't really be comparing them, and the 2023 batches were growing less or slower due to dealing with the aftereffects of covid or something.
Completely agreed - suggesting that this is a solution was a failure to think the next thought.
Nevertheless, if we had any idea how to actually, successfully do what Hinton suggested, even if we really wanted to? I'd feel a lot better about our prospects than I do right now.
Agreed, though we usually don't need to start with comprehensive understanding to start making things better. There are often institutional/organizational low-hanging-fruit choices that ameliorate particular harms at low or negative cost, which we nevertheless manage to just not do for many years or decades.
And most people do not know how to weigh costs against benefits, or to estimate either, or to evaluate the credibility of third party estimates of either. In many contexts, "Y has a cost, but will pay for itself in X years" (where X<10) is somehow not seen as a knockdown argument even in purely economic terms. Adding other positive non-economic effects sometimes, somehow makes Y look like a luxury good, even more out of reach, or somehow a scam.