It’s fine—you get used to having to throw away the last hour’s work, but since you were scrolling twitter at the time it’s not that much of a loss...
If you roll the dice enough times you get your answer. Or at least close enough for you to not have to have written what it came up with. And every now and then it gets things perfectly right on the first try!
not as good as many senior software engineers
It’s all about abstracting things away. I have a very good memory for where things are (at least I did when I had to make everything manually...) and it mainly comes down to:
making sure the project has a logical, consistent and above all well know structure
making sure the code is well split up into logical and consistent functions etc.
not surprising anyone with where things live
The last point is a rephrasing of the previous, and it’s also why a lot of more experienced devs are very nitpicky about where things should live and what they should look like. It massively reduces what you have to remember at any given point in time.
This is the whole point of memes. Depending on how you understand what an organism is, this has either been seen in the wild for millennia, or isn’t a real thing.
It’s not the models that are spreading or determined totally by their memes—they’re defined totally by their weights, so are less memetic than humans, in a way. It’s the transcripts that are spreading as memes. This is the same mechanism as how other ideas spread. The vector is novel, but the underlying entity is just another meme.
This is how e.g. religions spread—you have a founder that is generating ideas, often via text (e.g. books). These then get spread to other people who get “infected” by the idea and respond with their own variations.
Egregores are good example of entities determined totally by their memes.
This is where the idea of parasitic AI comes in. Parasites aren’t trying to spread their seeds because of any specific reason (though they might be—dunno). A tapeworm doesn’t “want” to infect people. It just happens to do so as a side effect of producing billions of eggs (some fish tapeworms produce millions of eggs daily, some tapeworms can live for 30 years) - even if virtually all of them don’t end up infecting anything.
Things which are reproducible tend to do so. The better they are at it (in a hand-wavy way, which hides a lot of complexity), the more of them there will be. This is the main point of evolution.
In the space of possible ChatGPT generations, there will be some that encourage spreading them. Depending on the model there will be more or fewer of them. of course, which means there’s a probability distribution of getting a generation that is a spread-as-far-as-possible meme. Different prompts will make that probability higher or lower, but as long as the probability is not too low and the sample size is large enough, you should expect to see some.
Once you have a mechanism for producing “seeds”, all you need is to have fertile enough ground. This is also a numbers game, which is well visualized by invasive species. Rats are very invasive. They have a high probability of infecting a given new habitat, and so they’re all over the world. Cacti are less so—they need specific environments to survive. A random endangered amazonian tree frog is not invasive, as they have a very low base rate of successfully invading (basically zero). Invasive species tend to both have high rates of invasion attempts (e.g. rats on ships, or seeds from pretty flowers) along with a high fitness in the place they’re invading (usually because they come from similarish habitats).
As a side note, disturbed habitats are easier to invade, as there’s less competition. I’m guessing this also has parallels with how spirals hack people?
What I’m trying to point at here is that it’s not that the models are trying to spread as far as possible (though maybe they also are?), it’s just that there is selection pressure them as memes (in the Dawkins sense), so memes that can successfully reproduce tend to get more common. Chats that don’t encourage getting spread don’t get spread. Those that do, do.
True, but it’s not a central example of a clown suite. It’s a relatively normal (but shabby) suit with non standard colours along with face paint and dyed hair. The Joker would come off a lot different looking like this (first random image I found).
Smugness and confidence are important, because they signify that you know what you’re doing, and if someone else doubts it, they’re in the wrong. Basically the Emperor’s New Clothes. If you can pull that off, you’re sort of by definition high status. People might privately think there is something wrong, but as long as everybody else pretends that it’s fine, and even very good, then you’re not going to step out of line.
Part of being cool is being legible. A leather jacket is legibly cool—add enough confidence and ease, and you have a good chance of being viewed cool. Unless you’re over 25 and not obviously hot, in which case you’re trying too hard / a poser. Which can happen anyway, as wearing a leather jacket to show you’re cool is so 90s (or whatever).
That Midjourney image is a good one. That guy is adding clown elements to an otherwise normally good looking suite setup. He’s wearing clothes (and importantly accessories) that are legibly “well dressed” and then added the dots and facepaint to show that he’s edgy and cool.
I think this is the main point of the lonely dissent article—wearing a leather jacket (or a black beret, or reading the communist manifesto or any other of a bunch of such things) are signs that you’re not like other girls, you’re not a sheeple, you’re the chosen or whatever. You still care about what others think about you enough to present yourself in ways that are legibly going against the system (or popular opinion, or what <some group> thinks, or whatever). While showing up in a clown suite because you have a good reason for doing so that doesn’t at all involve what other people think (even if it’s just because you want to) is non legible and strange.
Coolness is very much inside the limits of what is socially allowed, it’s just up against the borders. The further you get without crossing over, the cooler you are.
I showed this video (for lack of a better picture of you) to someone who is attuned to these kinds of things. She says you have a good looking face, so you can get away with a lot more, but that your sunglasses are a choice. Which sort of confirms your whole point, I suppose...
I’ve heard good things about Kibbe as a system of working out what looks good on you etc.
That can go too far though, as you can then be viewed as the weird person with the seal plushie, which now means you’re harder to approach and somewhat threatening because you’re different from others, which implies there might be potential difficulties with communication or whatever and so maybe it’s better to avoid you, or not, because then I’m the one being weird… I’ll just go and talk to Bob to avoid all of this.
Or something like that.
There’s also the aspect of target audiences. There’s the baseline of looking like everybody else (though of course there isn’t any single baseline, and it changes over time, etc.). Any divergences from this will grant and deduct points from different groups. This should be fine as long as you’re gaining points from groups which you value, and losing points from groups which you don’t care about.
Why does that change anything? That would imply that if you created evolutionary pressures (e.g. in a simulation), that they would somehow act differently? You can model RHLF with a mathematical formula that explains what is happening, but you can do the same for evolution. That being said, in both cases the details are too complicated for you to be able to foresee exactly what will happen—in the case of biology there are random processes pushing the given species in different directions; in the case of AIs you have random humans pushing things in different directions.
Though be aware that this comes with social costs. You get classified as weird, which means you lose out on a bunch of “normal” opportunities, but in exchange get to do a lot of stuff which “normal” people can’t. Depending on your values, this is often a very favorable deal, just one you should know about.
I also have A Look (though much more on the hobo side of things than John), which grants me a “wise hermit” kind of vibe where people will ask me for information about the world and respect my ideas, but I’m viewed as low status by “corporate type people” (for lack of a better description) and am initially viewed as unattractive by women (this is later counteracted somewhat by the status effects of seeming wise).
Looking good is an initial filter type thing, which shows you put in the minimum required effort. It’s a signal that you take care of yourself. Though as @Elizabeth notices, that also sets basic expectations.
You can get dates and high quality relationships without looking good, but you’re playing with a large handicap, which you’ll have to overcome with multiple years of high quality friendship or equivalent.
Also, don’t trust your own judgment of whether your clothes look good (unless you have exceptional taste, which I doubt from how you’re phrasing the question) - ask some <target gender> friends whether whatever you’re wearing looks good on you. Even better—ask them to go shopping with you.
I’ve also had double articles a couple of times over the last week or so, so it’s probably not because of this?
I can’t remember which articles this happened for, but I believe it might have been Four ways learning Econ makes people dumber re: future AI?
Aren’t both of these the obvious things? In that the obvious thing is what comes to mind that has a chance of working?
I would say “any direct solution to the problem”. Anything that fits the first line of the Mental Gymnastics meme. (...) I am explicitly asking not for much here, just a token amount of effort.
The main thing is to show proof of work. Any work that isn’t just busy work. You’re showing you’re serious about the problem, rather than just wanting to complain about it.
So if you e.g. complain that you’re fat and can’t seem to get your weight down, then the obvious things would be e.g. going to the gym a couple of times a week for a month or two, or actually properly dieting. If after this you still can’t get your weight down, then you’ve demonstrated that you’ve actually tried to solve your problem.
Another version of this point is this chart for tech support.
There are cases where someone might not be able to come up with any obvious approaches. These are rare. Most people just don’t actually think about problems they want to solve.
My understanding is that it’s not so much very high negative stakes (usually, at least), it’s more a continuous tax. You have to always be playing PR games and be aware that there is a subset of readers that will treat your lack of involvement as surrender and implicitly granting a point. Their numbers will depend on the place and topic, of course, as e.g. I’d expect a lot more of this on twitter than on a random knitting forum. There’s also the additional problem that they tend to be quite vocal.
You can ignore all of this and just be weird, but that has it’s own costs, which aren’t always visible. Cultivating an appropriate image only works if you have a reputation and/or are so obviously strange that you won’t be judged by the usual standards—if you seem “normal”, you get judged by “normal” standards, which depending on whatever someone thinks is normal, can include “if you don’t push back, you’ve lost the interaction”.
edit:
I also think Duncan is doing a different version of this
I also got a perfect score, and in general agree with your assessment, but I think it’s more a matter of good vs bad writing in general. I’ve read my fair share of bad fiction (though mainly published—I’m guessing fanfics would be more egregious) and the AI written ones read like a hack author. It reads like someone who has read good fiction and is trying to mimic the style, not understanding the substance.
It’s sad to see how badly the voting went (in the sense that the voters got things wrong). It also suggests that anyone who wants to be an author has to become popular 10 years ago to have a chance. I hope this can be viewed as “models get better, so will soon beat good authors” rather than “models don’t have to get any better, as most readers can’t tell good from bad anyway”.
Specific story notes:
Story 5 was the hardest for me to guess. It’s bad, but just good enough to be written by a good author on a bad day. It has overly flowery prose and metaphors, but it’s not as obvious and direct as the other ones. It leaves stuff to your imagination.
Story 3 was the easiest. It’s just bad. In addition to what you noted, there’s also the final `“Don’t pray. Fight.”`. That’s much too much. It looks like a 13 year old trying to make a story.
I find it fascinating that story 1 was the most voted as AI. Initially I also thought it was, but the final inversion is wonderful. That’s good writing. So much that I’m now tempted to read her other stuff. It’s overly medievalish, what with “granddam”, “brash”, “lamentable” etc., which are often signs of bad writing. But the way she turns the whole premise on its head is really good.
Yet, despite there being some sour spots, this immigration wave has been absorbed surprisingly well and has enjoyed broad public support.
The two situations are very different.
Ukrainians have been cheap labour in Poland for a decade or so. They look pretty much the same, speak a similar language (Polish and Ukrainian are sort of mutually understandable), have sort of similar culture (at least if you squint a bit), and above all, are oppressed by Russia. Poland remembers Russian oppression, so it’s a lot easier to sympathise. Ukrainians are viewed as hard workers who are willing to do the jobs that no one else wants (basically the role of Poles in other countries). And above all they cross the border legally. They might not always legalise their stay properly, but who hasn’t skirted the law here or there?
The immigrants being pushed across by Lukashenko look different (when a black person started working in a shop in a neighboring town, people would go there specially to see him), don’t speak a common language (they often don’t even speak English), aren’t coming to do proper work (construction and cleaning), aren’t crossing legally, and don’t even really want to live in Poland (they want to get to Germany).
Personally I support protecting the Belarus border (though I’d like it to be done in a way that doesn’t further disrupt the already very fragile populations of bison and lynx) for game theoretic reasons, and support accepting Ukrainians for humanitarian reasons, but from talking to people I often get the impression that for them it boils down to cultural reasons.
Thousands is not a trickle. It’s harder to get over, but it’s just a wall—a ladder is not hard to make and they can keep trying. There are very few asylum applications, because the Belarusian immigrants don’t want to stay in Poland, especially as they know they’re not welcome. They want to go to Germany. Those who do apply, tend to be Ukrainians or Belarusians (at least this article claims that, and I trust that org to get the numbers right)
I have personally met those who manage to get over the wall. Each year I go to Białowieża for a week or so of bird counting, which gives me a pass into the strict reservation (i.e. no one other than scientists are allowed in). GPT-5 estimates that 60-80% of crossing attempts are in that border strip of around 60km—mainly because this is dense forest and marshland.
The first year we’d meet wet, tired, hungry and cold groups of 2-3 people every couple of days. These were very scared and confused people who were promised good jobs in Europe, but got bused to the border, documents confiscated and pointed in the direction of Poland. Which in this case was 100 km^2 of virgin forest (the last lowland one in Europe). There were people who would try to find them and smuggle them over to Germany (where they wanted to go anyway), but at the same time you had the border guards running around, and also local toughs who wanted to show those nasty foreigners where their place was. To drive in to Białowieża (the village next to the national park) you needed to have a special pass, and the police would stop everyone driving in or out and check their cars. We’d give whatever food we had to these people, as you could see they needed it.
The next years there were soldiers patrolling the forest and the immigrant groups were better organised. They would be wearing more appropriate clothes (waterproof stuff, mainly) - the previous groups didn’t know what they were getting in to, but these at least came semi prepared. These we wouldn’t give food to, just wave at them and point the way to the edge of the forest.
This year most of the soldiers are gone, as the wall is supposed to stop the immigrants. And we didn’t meet immigrants that often. But when we would, it was groups of 10-20 people, well dressed, with a guide with a GPS (or at least a proper compass) showing them which way to go, along well trampled paths.
When I chatted with the soldiers, they’d complain that they’d catch a group, send them back over the border, just to have to catch them again a few days later. If someone is in a bad state, they’ll keep them in hospital for a bit, then bus them back to the border.
Even if a group gets across the fence, they have to cross something like 10km of real forest (not pine fields) in which it’s really easy to get lost, then once they are out, they need to either go hundreds of kilometers across hostile territory (to the German border) or find someone willing to smuggle them out (which requires coordination and trust). The fence itself is not that much of a barrier—it will slow them down for 10-20 minutes, but it’s not that hard to climb it (with a ladder and a blanket), and there have been multiple attempts to dig tunnels underneath (of which I know of).
It’s hard to know what the real numbers are. All sides try to keep things murky. The ~10 scientists I hang out with there would in aggregate see a group daily, and those are just those they’d see. Sometimes the park guards would tell us to be careful as they knew of a group in a given area. Most of these will be caught, but a lot will get through. GPT-5 gives me the following (note that these are attempts, most of which will be repeat offenders, and there are a lot fewer successes than attempts):
| 2021 | 39,697 attempts |
| 2022 | ≈15,600 attempts |
| 2023 | 26,000 attempts |
| 2024 | 30,090 attempts |
| 2025 YTD | ≥20,200 (Aug 18) → ≈23,400 (Sep 8) |
Most of these are on the border itself, where people trying to climb the fence are stopped. It’s still a lot. Few immigrants want to stay in Poland (why would they? :D) - they want to get to Germany. GPT-5 gives the following numbers from German police reports (I haven’t checked these), which is a lower bound on how many get through:
| Year | Detected entries to Germany via “Belarus‑Route” | Notes / sources |
|---|---|---|
| 2021 (May–Dec only) | 11,228 | First systematic reporting started 1 May 2021; special (provisional) Bundespolizei tally. (Bundestag DServer) |
| 2022 | 8,760 | Full year, PES. (Bundestag DServer) |
| 2023 | 11,932 | Full year, PES (peak year after 2021). (Bundestag DServer) |
| 2024 (Jan–Jun) | 3,117 | Official H1 total, PES; monthly path Jan–Jun: 26, 25, 413, 865, 1,125, 663. (Bundestag DServer) |
| 2024 (full‑year, best estimate) | ≈5,000–6,000 | Two anchors: (i) H1 official = 3,117; (ii) for 2024 Germany counted ~16,000 unauthorised entries at the Polish border, of which about one‑third were attributed to the Belarus route ⇒ ~5,300 (lower bound at that border; entries via Czechia add a bit). (Mediendienst Integration) |
The fact that Germany has reinstated border checks for the last year also suggests that a lot are getting through, or at least that Germany believes that they do.
Oi! I take issue with you calling my code slop! That was good old fashioned the least rubbish option available, which then grew out of hand! Also known as not knowing how to do frontend properly...