What's left to say | deadSimpleTech

Well, that was unpleasant.

My primary experience of the world at the moment seems to be exhaustion. I'm exhausted at the world and the people leading us, I'm exhausted at the way in which all of our social interactions seem to reduce to a kind of bleak superficiality (hence the networking article of a week-and-a-half ago, which was needlessly mean-spirited and for which I apologise). I'm exhausted by the fact that even people who should and claim to know better fall into that mode of interaction, and I'm exhausted by the fact that there increasingly seems to be no space in the world for those of us who have values beyond the dollar, the click, the like or the engagement statistics. I'm exhausted by the constant undercurrent of superficial anger in everything: not the kind of anger that leads to righting injustice, but the kind that leads to you quietly seething as you look for more things to make you angry. And I'm exhausted by the fact that attempting to be moderate, temperate and level-headed in this world seems to be a highway towards invisibility and irrelevance. Hell, even the great luminaries have to piledrive people to get eyes on their work: even in this blog I feel consistent pressure to channel anger to get readers, even though I'm not good at it and when I do it comes out with the kind of real venom that puts people off.

All this is to say that I think that perhaps we were lying to ourselves when we worried about AI and its social impacts, not because we said that the technology is bad and has negative consequences (it does), but because of what we told ourselves: we criticise the technology because it's an affront to human dignity and creativity and damages our ability to create art, write or otherwise be human. I think that that's probably why a lot of us got into writing criticism of LLMs, but while that may be the case, I can't help but see that many of the critics, including many luminaries in the field, seem to have prioritised hatred of LLMs over love for that which they threaten. They don't create much any more: they don't write things other than essays on LLMs, they don't make art or music, they don't even seem to write code or build systems much. And when they do, much of their output is sub-par. They've almost turned LLM criticism into an industry, producing page upon page of critique and making money from it. In the words of Contrapoints, lightly paraphrased, "they don't want to stop the damage that LLMs do, they want to endlessly critique LLMs". And in doing so, they commit many of the same sins that they accurately point out in the LLM world.

A society that would create both LLMs and the LLM criticism industry is one that's deeply sick in so many ways that it's difficult to count. It's not, in the end, that the AI creates slop that nobody wanted in the first place but that we've now developed an addiction to. No, we were in the business of consuming and producing slop for a very long time. We loved the stuff: business reports that nobody read, infographics that looked pretty (or even ones that didn't look very pretty and had downright sloppy design), youtube video essays that were just a person reading out a wikipedia article, SEO-calibrated business blogs that are completely vacuous and an endless supply of social media content. As much as we might say that we want challenging content that expands our horizons, broadens our minds and helps us grow, when faced with the choice between consuming something difficult or sticking our faces in the trough, we tend to stick our faces in the trough. It's for this reason, perhaps, that so many LLM critics, in turning their well-deserved ire upon the technology, have fallen into the production of slop themselves: they simply tend to do it manually rather than programmatically. Perhaps this was inevitable: it certainly seems in my experience that it's basically impossible to succeed at all in this society without making some slop, and even people who say they dislike slop, by their revealed preferences, will consume it over something more challenging.

Just look at the pressures on your average content creator. Even the phrase "content creator" speaks volumes: we aren't artists, writers or musicians, but people whose job it is to create a steady stream of stuff that captures the attention of the people who consume it so that our captive audiences grow and some of them send us money ¹. In that kind of situation, and with the kinds of margins that we work on, there is a very strong pressure to create a lot of stuff (which necessarily means sacrifices in depth or quality: I have maybe one really in-depth essay in me a week, along with maybe something lighter if I'm up for it and not too stressed), and to create stuff which is easily digestible, hooks people in and reinforces their beliefs rather than challenging them. Whatever people might say their behaviours are, that's what you have to produce to get people to give you money. The revealed preference of your average reader is for slop, and so that, increasingly, is what the people who want to survive end up producing. Note that AI has nothing to do with this: it's pretty easy to create LLM criticism that, while manually produced, is fundamentally just as much slop as anything that comes out of Claude or ChatGPT.

You can see the patterns: increasingly hardline positions on matters of fact that don't square with reality (the idea that LLMs are theoretically incapable of producing things that we see them produce with our own eyes), emotionally-charged language that aims to stir up rage, and above all, lots of content. It repeats itself, it's shallow and it makes the same few points over and over again, but my word do you see a lot of it, whether it's a lot of small posts or comparatively fewer very long ones. Everything is aimed to capture human attention and get people hooked on outrage: the question of whether a person might benefit by reading it or whether a virtue might be developed falls by the wayside. If that isn't slop, what is?

Running a consultancy or a business these days is also, by and large, a matter of producing slop. You need to start with a website with good marketing copy and good SEO: both of these are fundamentally the production of slop. Marketing copy, after all, is meant first and foremost to elicit an emotional reaction that gets people to do the thing that you want them to do, and SEO doesn't even do that: it aims to trick one kind of machine or another into surfacing your work so that humans can read your marketing copy. Even at its best-crafted, this is slop: it has an effect, not by quality, but by sheer volume of stuff that crosses the bar that you can put out. Having done that, you need to go to networking events, which are more or less social slop inasmuch as they let you develop a large number of connections that resemble a business relationship from the outside but that have little of the substance: again, the goal here is volume of interactions in the hope that one of them might turn into something more positive. And then of course, there's LinkedIn. You might, if you're naive, avoid the platform because its total economic value is slim: very little of actual good comes of it, the connections are of tenuous value at best and for the most part you just get to see the worst people in the world say horrible things in public with not a shred of shame. Alas, this is a mistake, because the social expectation that you be on LinkedIn means that if you aren't on there, nobody will take you seriously. This has gotten to the point where, contrary to what one might naturally expect, in-person meetings have become a prelude to LinkedIn connections.

And that's to say nothing of much of the software that actually gets written today. It's my view that more or less every application that we have a real use for has been mostly written: at most you might snap a couple of different existing pieces together and write a thin front-end on top of them to get what you want to do done. There are exceptions, but for the most part, we write far more software than we need to. This excess software is also slop: it's written for reasons of status, Not-Invented-Here syndrome or Veblenian entrepreneurship. The quality is naturally mediocre and whether it's been successful or not is judged largely on volume: it's software as commodity, for which we pay $2000/short tonne.

If we're drowning in slop, then, it's because on some level the vast bulk of the population likes it and has done for some time. This is independent of LLMs (which are mostly damaging because of just how well they pander to this vice): it's been a vice for a long while and has been getting steadily worse for at least the last few decades. If we're to deal with all this, then, it isn't just enough to oppose AI: we need to kick our addiction to slop and work to develop the virtues that slop consumption damages. Otherwise, if we oppose the constant use of LLM tools but don't challenge these other tendencies in ourselves, we'll simply replace one version of slop with another.

The end result is as we see it; a deep sickness of the soul where, whenever we're given the option, we choose ease, to have our existing beliefs confirmed, to be made angry in a way that feels good, to feel as though we're already all that we need to be rather than making the effort to grow. It's a world where truth doesn't matter, beauty is reduced to whatever's the most easily accessible and where nobody ever has to learn anything to access the best of what we can do. It's an intellectually impoverished world, where we consume neatly packaged little bits of insight slop as a substitute for the hard work of real learning, and if anything is at all difficult or not immediately graspable, it's discarded. It's a world in the process of slowly dying, where even opposition to this state of affairs tends to partake of these vices because otherwise it'll be ignored, and in the process neuters itself. It's the hell of TS Eliot's Hollow Men writ large.

The anti-slop virtues

So, what virtues have we lost? Truth and clarity of thought were the first casualties: in a world where attention is fleeting, the feeling of insight has to come immediately. You have to feel the truth of a statement as soon as you see it: anything that resists or requires you to put in work to understand falls by the wayside. And so we gave up thinking carefully and seriously: if something sounds true, it is true even if it turns out to not be the case. We can say what we like so long as it sounds true to our audience, however much it might not actually comport to the state of the world. Heaven forfend that you ask people to not believe false things.

Along with that we lost complexity, fair-mindedness and the ability to balance competing considerations. When something is important these days, it's the only important thing. Two things can never be both important and carefully traded off against each other: that is a level of complexity that slop simply does not admit. One thing, and only one thing, must be important.

We also lost intensity and depth of feeling, it seems: our emotions these days, though they might be intense, are fundamentally shallow. We cannot grieve but performatively, we cannot show joy without linking it to B2B sales, and our anger, rather than driving us towards righting injustice, simply leads us to consume more content. The whole complex texture of human emotional experiences has been flattened into slop. Raskolnikov in the 21st century isn't consumed with guilt over killing an innocent woman just because he could: he just goes and shoots Charlie Kirk, which, while it might make for great content, would be a disappointing novel.

With that list, we might note that the final aesthetic virtue of unity has also been lost. Good art coheres with itself: it has a shape to it that holds together as one thing. In the world of slop, that isn't viable: content cannot exist to support something else, but each individual element must draw attention to itself. We drown in a sea of things that are individually pretty but blur into noise.

And of course, who can forget prudence, temperance and moderation? Self-control is no virtue in the world of slop, and unrestrained emotion and consumption in their ugliest possible forms appear to be the order of the day. Restraint of any kind is an insult, an attempt to sabotage, something that gets in the way.

When we lose these virtues we not only become morally worse, we lose a large part of that which allows us to act in the world: we lose our agency and our creativity, and saturated by slop as our minds are, our ability to look for something better becomes sharply limited. We become passive, reactive people, consuming content and reacting in shallow ways to what is presented to us, unable or unwilling to actually try and change things beyond consuming more content or, at the outside limit, creating content of our own that will somehow change the world if it's consumed enough. And so the world becomes a little flatter, a little greyer, a little more exhausted, a little closer to a death by dissolution.

I love not the sword for its sharpness

I am a critic: critical analysis of the world is what I do best, and this is what I try to do in my writing. I am also, however, practically-minded and like to see myself make the lives of real people better, hence my work in consulting. Both of these fields have been consumed by slop to the point where, in order to compete in either one, I'd have to produce it myself, whether it's manually in the form of increasingly frequent, increasingly sloppy essays, or whether it's in the form of LLM-generated tech that's generally a terrible idea but that looks cool to the people in the market that actually have money at the moment. I've dabbled in this both ways to an extent, but I simply can't make it stick: neither the writing nor the code sit at all well with me. I've consequently spent a large part of the last few weeks in a bit of a depressive hole, completely unable to see how I'm going to survive until things improve or some kind of opportunity eventuates for me: it really feels as though neither writing, nor tech, nor academia have much space for someone who doesn't want to produce slop at the moment.

What's kept me going through this has been, of all things, my statistics students. I've been teaching a couple of people working in professional fields statistics lately, beginning from the foundations in set theory and the construction of a probability measure and working towards the important statistical tests from first principles. I've still not got quite enough work coming in for real stability, but above everything else this has been good for my soul. I've not had to simplify or sloppify anything: in fact, this is probably about the least "business-friendly" presentation of the material that I've ever seen, but my students seem to appreciate it even more for that. And I can see them learning and growing: they're very quickly picking up a lot of the values that I talked about above. Precision in thinking, accuracy, the ability to reason about how likely some things are or aren't to be true, the ability to balance complexity... well, admittedly they were all very smart and good at that to start with, but seeing them gain the confidence that comes with not only knowing something but knowing that you know it has helped me feel that what I do is worthwhile. It's the direct opposite of slop logic, and more than my consulting and tech work, or even my writing, it lets me feel as though I'm defending the human values that I care about against the onslaught of slop, both LLM-generated and otherwise.

What does this mean for our current situation and for LLM criticism? If we're to get out of this horrific situation, we need to stop with the endless slop logic, develop the virtues in ourselves that will let us let go of our attachment to slop and actually strive positively forward for the values and the virtues we want: humanity, truth, clarity, beauty, Quality (in the Pirsig sense). It's not enough to criticise the latest attack on these things that LLMs represent: we need to actively cultivate the virtues that curtail slop and make us want something better. We have to learn the hard things, develop real mastery and potentially even read a difficult book every so often: otherwise, much as we might criticise with the best intentions, we will end up simply replacing one form of slop with another.

It's difficult work to do this in current circumstances; there is little money in non-slop content and as "do a revolution in which we purge all the slop from the earth" is not a practical way to solve any problems (it's downright eschatological) we'll likely find ourselves interacting with one form of slop or another in our lives. Still, it's important to carve out some space in the world where you can do work that isn't slop: it's vital for your mental health, and for keeping you honest when you do have to compromise on one thing or another in the end. You might have to produce slop in most places, but you have to have at least one place where you can do work to your own standards.

As for the critics, I don't wish to criticise them unreservedly: they have done important work, and in doing so have often built strong media networks and strong audiences. They do, however, often sit in positions of relative privilege as a result of this, or flowing into it: many of them are academics at prestigious institutions or have worked at the high-and-mighty institutions of the world (even if you got fired from Google, you still worked at Google). They have a responsibility, then, not to simply sit and endlessly critique the tools for their own gratification, but to make life better for those of us who don't have their advantages and can't make a living from endless critique. I'm not sure how that'd work: perhaps they could share their networks with the rest of us, though I'm not holding my breath. But just as everyone has a duty to advance their vision of the world rather than simply endlessly critique the current one for clicks, so do they.

Also, if you'd like to learn the fundamentals of statistics and statistical thinking from me, please get in touch! I've already had quite a few happy clients, and it's one of those rare things that will make you both happier and more effective in your job and life.