defgeneric

We're going to see the battle intensify here because "spikes" of ASI are emerging that can't be ignored. Models are now better than humans in certain domains or for certain tasks, which means capital as a moat is being eroded. This is why you see people like Jamie Dimon sounding the alarm. Most of the talk until now has been about how the models would be a serious problem for labor, but if that was true how could it not also be an issue for capital?

Yes, and given the nature of the current administration, whose actors are not inclined to see themselves as independent competing capitals among others, but rather as privileged capitals, and therefore more inclined to move towards taking an interest in the process of enclosure, ensuring it includes them, the hope seems to lie with companies at the hardware layer who have an interest in seeing the diffusion of intelligence play out freely at all levels of society.

The labs have to be told NO--the problem they're dealing with, that model outputs give the game away, and that in turn the distiller becomes the distilled, is a fundamental problem they have to figure out how to deal with without going to the State.

This is where we need the hardware companies and neoclouds to start speaking up. The labs want to elevate matters from the level of civil society (basically, competing firms) to the State (enclosure), and as always, in the name of security. But other actors in the same ecosystem have strictly opposed interests here, and are equally if not more credible as far as the State is concerned. If players like Nebius, Baseten, Fireworks, etc. among many others including obviously Nvidia, Dell, AMD, and so on don't get ahead of this they will be sacrificing trillions.

Like all forms of eschatology it's not actually concerned with the future to come--it's real aim is to have influence now.

It's just a form of rhetoric.

To dismiss them, saying, "ah ha! they're making the same error as all failed prophets!" misses the point.

What they are really talking about is AI governance.

Approximate as a limit and not surpass, where the hidden hope actually seems to be that they don't surpass. There are lots of variations floating around, from simple metaphors like "LLM as librarian that speaks to you," to even Sutton's remark that rather than being a case where we've taken the "bitter lesson" to heart, LLMs may be a yet another case where we're limited by it.

It's a form of denial. We're getting another "de-thronement of man" on the order of Copernicus and Darwin. Some get excited, others turn away in horror. Negation is the outward expression of the desire to keep human intelligence wrapped in its mystical veil.

One popular idea is that these systems will asymptotically approximate human intelligence because they're trained on mostly human-written texts. Not only is that untrue, it's also directly contradicted by our experience with previous RL-systems, where they seem to breeze right by human ability without even the slightest hiccup.

A justification for this would be: states choose these providers to accurately present building codes (keeping up with the right versions/revisions is a lot of work, etc) because there's risk in simply publishing the PDFs and leaving it to basically anyone to curate. But the providers (like ICC) also tend to lock these deals in with state laws (not sure if they lobby for it, but I would imagine they do). When you think about who needs to have access to the correct and current building code or electric code or whatever for regular reference, it's really your lawyer, plumber, electrician, architect, builder, etc. and in that context $170 or $130 isn't a bad deal.

This is why Google is pushing SEOs to get their clients to codify and publish their domain expertise: while it gives them a way to filter signal from noise/slop right now (supposedly helping to "improve search experiences"), it also simultaneously extracts that experience into a consumable form for later training.

They really do want to know the ins-and-outs of the HVAC service business, for example, because they hope their agents will be handling it in a few years.

In many cases the really key idea that transforms the overall system design comes from working closely on the specific implementation details. Maybe you don't redesign the system this time, but you saw how you might do it, and you get ideas about how to do it the next time. The craft involves a back-and-forth between different levels of abstraction, and cutting that link does feel like we're sacrificing something.

Earlier automation targeted physical and manual labor, where a worker’s identity was at least partly separable from the output. A welder is not the weld.

It comes through even worse in this later sentence, which at the very least tells you the writer has never met a welder.

In case it helps anyone, in some minor cases I was able to recover and continue with /rewind.

There was also a period around the mid-2010s where I had the strong impression that lots of younger ambitious devs were fanatically promoting rust against C's undefined behavior mostly because it gave them a way to differentiate themselves from older seniors within organizations. (And I say this not as an old C diehard, but as someone who watched more than one colleague position himself as the 'rust guy'.)

It could be a symptom of how fragmented workflows are, which itself seems to be due to providers adding friction to guard against being integrated away by some larger platform.

You mention the technical aspect (readability) and others have suggested the aesthetic, but you could also look at it as a form of rhetoric. I'm not sure it's really effective because it sort of grates on the ear for anyone over 35, but maybe there's a point in distinguishing itself from AI sloptext.

Incidentally, millenials also used the "no caps" style but mainly for "marginalia" (at most paragraph-length notes, observations), while for older generations it was almost always associated with a modernist aesthetic and thus appeared primarily in functional or environmental text (restaurant menus, signage, your business card, bloomingdales, etc.). It may be interesting to note that the inverse ALL CAPS style conveyed modernity in the last tech revolution (the evolution of the Microsoft logo, for example).

The development of QM was so closely connected to experiments that it's highly unlikely, even despite some of the experiments having been performed prior to 1900.

Special relativity however seems possible.

I suspect the move back to pen-and-paper exams is being resisted by the teachers. It shouldn't be that hard though--when the workload became to great, most of my own professors would offload part of the grading task to TAs and grad students.

It does seem like in-person pen-and-paper exams would hold the line pretty firmly with respect to competence. It's a simple solution and I haven't heard any good arguments against it.

Also known as the "Gell-Mann amnesia effect" [1].

[1] https://en.wikipedia.org/wiki/Gell-Mann_amnesia_effect

I'm seeing this reaction a lot from younger people (say, roughly under 25). And it's a shame this new suspicion has now translated into a prohibition on the use of dashes.

How do I belong? How do I make my community a better place? How do I build wealth for the people I love?

What remains after is something like the social status games of the aristocratic class, which I suspect is why there's a race to accumulate as much as possible now before the means to do so evaporate.

Note we currently live in the most surveilled state in history.

It's garbage opinions like this that makes PG so tiring. The superficial air of reasonableness makes it attractive to younger SF tech people who haven't experienced the context out of which these arguments arose and have no idea who he's plagiarizing/channeling. (For starters, the distinction between wealth and money/capital goes back at least to the 17th century.) For those who are more interested in being the "next unicorn" than engaging seriously with ideas, his little "essays" serve as kind of armor--we don't have to think about that problem because PG wrote about it!

If it's any consolation, these companies paying for ads on a competitor's brand name are probably paying through the nose to get clicks that only bounce. IF it's worth it at all, it's probably temporary. It's an indicator that market share is still up for grabs.

This was my first thought too. The last generation of activity tracking, while still dystopian, was a little different at least in that it was mainly statistical. So action-wise, it might point managers at "potential problems," but doesn't make its way into a performance review (e.g. "your mouse only traveled 81.72 screen-miles this quarter, 2 standard deviations below the mean, while you also scored the lowest on number of keystrokes with vscode as the active window..."). If a manager really wanted to summarize exactly what was done they had to spend an almost equal amount of time to watch. To some degree, this alleviates that.

Why would software be qualitatively different from all other forms of automation that came before? And suppose, for the sake of argument, that software is fully automated at some point--what then happens to the firm?

Tao is great in lots of ways--obviously as a first-rate mathematician, but also as an educator and an ambassador of mathematics to the general public. It's cool to see him thinking along these lines, but if anyone is really interested in where this line of thinking goes, it's basically the problem of modernity. Just about everyone in the humanities is fully aware of this already. It emerged vaguely around the 1830s and basically became the major subject of the humanities--in one way or another--ever since. Marshall Berman's book All That Is Solid Melts Into Air is good intro. I would expect that if you take Tao's specific line of thinking here (society beginning gas-like, interacting particles, then clumping together at various levels of abstraction, and interacting up and down the levels, etc) you get into all sorts of issues that were debated endlessly a very long time ago. But as a quick, temporary prism for looking at the world it's fine I guess. As a symptom, it makes one think something else might be going on when a very famous mathematician is suddenly now rediscovering modernity--perhaps things become more clear the more they break apart. One might even go so far as to say Tao's apparent ignorance of the issue of modernity has something to do with specialization, i.e. is a symptom of modernity itself.

PS--to add one thing as a criticism, the "retreat" into "grassroots groups" has also long been viewed as a false solution to the problem. Politically, this "solution" emerged in various forms: the 19th century's utopian socialism (especially in the US!), late 19th-early 20th century syndicalism, 1960s communes and "turn-on, tune-in, drop-out," up to now with the stubborn idea that communal living is somehow "revolutionary" and various other guises. It's there in less "radical" forms too, like when liberals say we just need to restart the bowling leagues. It's fine as an individual respite, but will never really get at the problem, not least because there are many other (and better!) ways of getting some "relief".

A clerk who once used a physical filing system but now uses a computer may naturally begin to bring the filing system to a higher level of abstraction. Think filing clerk -> database administrator -> relational algebra...

E.g. there would be enormous difficulty in replacing the Dewey Decimal System with something else, if only due to its physical inertia, but with a computer system a curious clerk can invent an alternative categorization and retrieval system which inevitably touches on mathematical topics.

I've seen the same and it seems due to people moving down the stack in response to LLMs being able to code at a junior-ish level and everything that entails.

The physicality of having to actually do things in the real world slows things down to the rate at which our brains actually learn. The "vibe coding" loop is too fast to learn anything, and ends up teaching your brain to avoid the friction of learning.

This is exactly the problem, but there's still a sweet spot where you can get quickly up to speed on a technical areas adjacent to your specialty and not have small gaps in your own knowledge hold you back from the main task. I was quickly able to do some signal processing for underwater acoustics in C, for example, and don't really plan to become highly proficient in it. I was able to get something workable and move on to other tasks while still getting an idea of what was involved if I ever wanted to come back to it. In the past I would have just read a bunch of existing code.

A lot of the boosterism seems to come from those who never had the ability in the first place, and never really will, but can now hack a demo together a little faster than before. But I'm mostly concerned about those going through school who don't even realize they're undermining themselves by reaching for AI so quickly.