Login
You're viewing the social.sp-codes.de public feed.
  • Oct 15, 2026, 5:30 PM

    It's also very frustrating to constantly have my opposition to AI be taken as an opposition to abstraction, or to automation, and a desire to do every little thing by hand. But it's the opposite. For a whole bunch of reasons, AI anti-abstracts. It adversarially creates oceans of duplication to process. It creates a situation where the AI user gets to set the terms of some ecosystem with synthetic productivity but then everyone else needs an AI to figure out what's going on in that new ecosystem.

    💬 1🔄 86⭐ 124

Replies

  • Oct 15, 2026, 5:32 PM

    Abstraction is leverage. Automation is the amplification of power. Abstraction is democratization. Automation repeats a duplicative process so you can stop thinking about it and focus on something higher level. Abstraction elides so architecture can expand.

    AI is anti-leverage. Where an abstraction will elide duplication with semantic compression, AI steamrolls the duplication by copying and pasting faster than a human can.

    💬 0🔄 26⭐ 29
  • Oct 15, 2026, 5:35 PM

    Linus is saying the quiet part out loud: AI-native projects are environments where the key ingredient is _capital_. Abstraction can be performed by a quiet appreciation. True automation can be disruptive.

    AI-based architectures (again, assuming that the AI works reliably, which is a highly contingent proposition) translate all that stuff into capital flows. A "machine speed" project requires _money_ to participate in, regardless of their complexity.

    💬 0🔄 20⭐ 34
  • Oct 15, 2026, 5:38 PM

    Right now it looks like AI can do "disruption" because you can apparently do things that require a team, as an individual, by taking advantage of highly subsidized token rates, or (at best) taking advantage of highly subsidized training regimes and stolen training data. But those rates will go up and while you'll still be able to nominally command those resources as an individual, doing so will still require the capital flow required to hire a team.

    💬 0🔄 5⭐ 29
  • Oct 15, 2026, 5:40 PM

    AI does not disrupt. AI entrenches. AI turns money into more money; AI turns investment into dependency. We are living in a funhouse mirror of that right now where the combination of overly-plausible demos and first-one's-free incentives are swamping the actual balance-sheet math. But the economics require that this stabilize, that the capital flows reverse direction. The wave has come in and we are just starting to feel the undertow.

    💬 0🔄 15⭐ 52
  • Oct 15, 2026, 6:12 PM

    This is an attempt to address the growing number of "oh, so you don't like automation when it comes for *you*? checkmate, programmers" articles out there, but there's so much more to say. Overall I think this idea that computer programmers just destroyed jobs is largely false. One of the weird corollaries of the Productivity Paradox is that we mostly *did* expand possibilities and make things better and more fun.

    (Until all the Gig Economy stuff. But that's still a minority of tech history.)

    💬 1🔄 2⭐ 17
  • Oct 15, 2026, 6:16 PM

    @glyph I mean, yes and no. Reading about the final days of the typesetter trade here in Denmark is decidedly painful for a programmer with Luddite sympathies.

    💬 0🔄 1⭐ 6
  • Oct 15, 2026, 6:17 PM

    @datarama "mostly" is load-bearing. Industrial control systems also have some, uh, interesting labor relations implications.

    💬 0🔄 0⭐ 2
  • Oct 15, 2026, 6:18 PM

    @datarama But the simplistic narrative that Computers Are Just A Tool Of Capital is quantitatively incorrect overall. When you weigh the destruction of the typesetting trade against the growth of desktop publishing and the artistic possibilities of online media, the simple story of "programmers made typesetters poor so they could get rich off the difference" is a wildly reductive take on the whole thing.

    💬 0🔄 0⭐ 4
  • Oct 15, 2026, 6:26 PM

    @glyph @datarama Overall I think it's wrong to deem automation as anti-social when the processes being automated have rigorous objective metrics for whether they're being done right, and where a human operator is necessarily going to have a higher error rate. The entire perspective comes from a perverse "work ethic" worldview where it's Good for humans to be doing labor. Reducing the overall amount of labor we need is a good thing!

    Where automation is harmful is where it's replacing complex judgements, interpretations that have a lot of hidden context, and roles where there is an imperative for someone to be able to say no on ethical grounds with sloppy heuristics.

    These principles predate "AI".

    💬 0🔄 1⭐ 4
  • Oct 15, 2026, 6:42 PM

    @dalias @datarama there's a thing buried in here too where our society refuses to properly invest in retraining and employment disruption insurance programs. if you invent a miserable-shit-shoveling machine that perfectly and immaculately does the work of a miserable-shit-shoveler, immediately freeing them to do something other than miserably shoveling shit for a job, and in the course of doing this you make them homeless and despised, you can't really fix the *machine* to address that

    💬 1🔄 2⭐ 4
  • Oct 15, 2026, 6:43 PM

    @dalias @datarama everybody who does a properly obsoleted job SHOULD be given a substantial bonus, emotional hazard pay for having to reformulate their sense of self, and maybe a parade to thank them for their service for getting society to that point, instead of quiet desperate irrelevance and a probable ignominious death. but I don't know if it's quite fair to heap all that at *programmers'* feet

    💬 0🔄 2⭐ 6
  • Oct 15, 2026, 7:03 PM

    @glyph @dalias When I was in my late teens, if I went to a public park there'd sometimes be a bunch of people mowing the grounds, doing simple weeding work and the like. Most of them had mild-to-moderate cognitive disabilities, and this sort of "menial" work was the only thing left for them to do.

    Now they've been replaced with robot mowers - but many of them, now in their senior years, still hang around the parks, picking up bottles to collect the bottle return fee, and sometimes just cleaning up trash if they find it. They're not economically desperate or homeless (and you don't make *that* much from a day's worth of bottle returns) - virtually all of them are on disability pensions, which is a livable income here. I've talked to many of them (as I also did when they were employed and I was some young delinquent smoking weed in the park, so some of them instantly recognize the older and at least outwardly more respectable me). They do it because they like to feel useful. That at least they're contributing *something* to the community. They don't like to feel "useless".

    And honestly, this is *the same thing* I feel when I'm staring down the barrel of unemployment.

    Now, I should mention that I live in a culture with a social-democratic work ethic, which is perhaps the only thing that's even harsher than a protestant one. But it generally seems to me that many people genuinely like feeling that they're contributing something to the fellowship, and often feel bad if they're deprived of a way to do so.

    I'm not saying we should take Enoch's hammer to the robot mowers! But I'm saying that something that *to me* looks like miserable shit-shoveling might well be the thing that makes someone else feel like there's at least a small thing they can do to keep the community running. And if the last 30 years have taught me anything, it's that programmers are not necessarily the right people to make these judgment calls.

    And certainly, programmers organized in privately-owned corporations aren't.

    💬 1🔄 4⭐ 8
  • Oct 15, 2026, 7:18 PM

    @glyph @dalias Now, humans are not tools and are not made for any particular purpose, so "useless" and "useful" are not really terms that make sense to apply to people. And much as I struggle with the now-near-constant intrusive thought of viewing myself as "soon useless" because LLMs are being deployed to do things that used to be meaningful parts of my work, it is not a term I would ever apply to anyone else. So obviously there's some cognitive and emotional distortion going on here, for reasons that are probably as much personal as anything else.

    But I do think the wish to be "useful" - in the sense of providing some kind of contribution to the well-being of the community, however small - is something that sticks deeper than culture - and therefore, also deeper than the constructs of wage labour and productivity.

    Personally, aside from *seriously* not liking the thought of no longer being able to meaningfully contribute - regardless of whether I was economically comfortable on an adequate UBI - I find a very strong motivation in doing a job well. I enjoy writing good code, I take a craftsman's pride in it. This is another reason why LLMs feel like a straight-up insult - and why I was already extremely frustrated with my trade before they showed up, because the entire incentive structure has long been arrayed against the struggle for quality.

    💬 0🔄 0⭐ 4
  • Oct 15, 2026, 6:28 PM

    Also this is a big argument in the alternative. For technical reasons related to context limits and training data, none of this shit actually works well enough to replace software at any meaningful scale. To paraphrase @pluralistic by way of Jensen Huang: AI will not take your job. A boss using AI to fake their quarterly KPIs will take your job.

    💬 0🔄 9⭐ 22
  • Oct 15, 2026, 7:29 PM

    @glyph I've been saying this for years ever since "it just speeds up the boilerplate" first ruffled my feathers and I had to sit and examine why.

    If boilerplate is somehow prohibitively uncomfortable to write, that's a pain signal here to tell you something. You're missing an abstraction! Just like touching a hot stove should signal you to not do the thing. Choosing to just do the boilerplate faster with stolen training data, even if it's a task the bot can easily do correctly, is like shooting up morphine instead of moving your hand off the stove. And nearly all LLM apologia is like this!

    If nothing else, I have a far more fundamental respect for friction* and pain signals than I did five years ago.

    💬 1🔄 79⭐ 179
  • Oct 16, 2026, 5:55 PM

    @miss_rodent @MaddieM4 I feel like this is a bit more complicated than you’re making it sound; on a *large* project, this is absolutely true, but a lot of software work involves tons of smaller projects, each with its own codebase, that is rarely revisited. each unit of abstraction (package, service, host, node, app) creates a level of non-negotiable boilerplate overhead, often in the form of configuration or data, not even code

    💬 0🔄 0⭐ 4
  • Oct 16, 2026, 5:55 PM

    @miss_rodent @MaddieM4 the point still applies but it’s often a big challenge to figure out how to tackle those sorts of abstraction problems especially when each thing really is a *little* bit different in a way which is difficult to harmonize

    💬 0🔄 0⭐ 4
  • Oct 16, 2026, 5:55 PM

    @miss_rodent @MaddieM4 AI is still a siren song that is going to make the problem worse rather than better but I absolutely understand where the temptation is coming from

    💬 0🔄 0⭐ 4
  • Oct 16, 2026, 7:03 PM

    @miss_rodent @MaddieM4 once again there's a very interesting tension to explore on this topic but LLMs just pull in the least productive possible direction and drown out the more useful ecosystem conversation we could be having

    💬 0🔄 0⭐ 3
  • Oct 16, 2026, 7:04 PM

    @miss_rodent @MaddieM4 I could give an impassioned defense of both big-library and small-library worlds and do a whole giant thesis/antithesis/synthesis thing about it. and I have literally had conversations and collective activities about this problem pre-empted by needing to react to LLM nonsense to prevent specific and acute damage elsewhere

    💬 0🔄 2⭐ 5
  • Oct 16, 2026, 7:16 PM

    @miss_rodent @MaddieM4 I … literally do not have the time to get into it. I need to do other stuff today! Augh!

    At least only like … 25% of it is reacting to LLM-caused problems :)

    💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 1
  • 💬 0🔄 0⭐ 2
  • Oct 16, 2026, 7:30 PM

    @miss_rodent @MaddieM4 I don't think I've ever done a big write-up of what I think is the most significant issue, which is that the whole debate over big vs. small libraries and short vs. long dependency lists is mostly a red herring as compared to the debate that we SHOULD be having, which is drawing an explicit social relationship graph over the transitive closure of all dependencies

    💬 0🔄 2⭐ 4
  • Oct 16, 2026, 7:31 PM

    @miss_rodent @MaddieM4 part of the difficulty of that conversation is that open source is explicitly designed to *stop* that conversation from happening, because it is annoying and expensive. you just "adopt a dependency" or "don't", you don't "establish a vendor relationship with the open source upstream".

    💬 0🔄 2⭐ 4
  • Oct 16, 2026, 7:32 PM

    @miss_rodent @MaddieM4 every attempt to fix funding (tidelift, thanks.dev, gratipay, even github sponsors kinda) is a transparent overlay where a third-party vendor will establish this set of relationships implicitly on your behalf

    💬 1🔄 1⭐ 3