gwern
So, most LLM RL training would be expected to exacerbate this issue?
I think that it’s a pity if people write off my SD page because they failed to understand the meaningful illustration I put effort into creating and didn’t, say, check the alt text to see if they were missing something or wonder why such an unusual website would have “AI slop”; and I agree that this may be a case of “things you can’t countersignal”.
However, I refuse to submit to the tyranny of the lowest common denominator and dumb down my writings or illustrations. I don’t usually write for such readers, and I definitely do not write my Gene Wolfe essays for them!
So unless people can point to something actually bad about the illustration, which makes it fail to satisfy my intent—as opposed to something bad about the readers like being dumb and ignorant and writing it off as “AI slop” when it’s not—then I decline to change it.
The image borders are exposed, and the white fade does not work. Compare to the fully functional light mode:
To be honest, I don’t like the light-mode example either. I think it’s bad to have your text visibly overlapping like that. (If I made an image which I had put
.float-right .outline-noton to get the same effect on gwernnet and I saw your ‘good’ white-mode version, I would be immediately complaining to Obormot about his CSS being busted.) So in this example, isn’t a lot of the problem that the UI elements are overlapping so the text is spilling over onto the scales and blocking the wires etc, and the dark-mode merely exacerbates the problem and fixing the core problem would also fix the dark-mode issue?
FWIW, my impression is that while the gwern.net dark mode was a ton of work to create, due to requiring a total refactoring of the CSS and deal with the flash-of-white issue and figure out the correct three-mode behavior, and create tooling like the color conversion script or needing to create InvertOrNot.com for acceptable images, or tweak the syntax highlighting colors to look right in dark-mode… But once we finished paying all of that, the maintenance burden has been relatively small. I have to patch up the occasional image not inverting right and we have to occasionally special case some Wikipedia popups CSS, but it’s on net less effort than much of the website (eg. Apple device support, or the local archive system, are perennial hassles).
The gwernnet dark-mode is pretty good engineering work overall. We should’ve done a design-graveyard writeup to explain not just how it works (non-obvious) but the problems with the more obvious approaches and what problems we had etc. I fear it might be too late to write it now even if we had the time...
I don’t believe it is “AI slop”, much less that it is “pretty nasty”. I consider AI slop to be low-meaning and low-effort generative media which adds little or nothing to the experience
I assume you are referring to the German Expressionism, alluding to Nosferatu (which is highly relevant for at least two reasons), image illustrating the narrator’s childhood iceskating in a New England Protestant town in decline due to Dracula taking it over; I generated it in MJ after cracking SD, to sum up the horrifying reality of my solution. I put several hours of thought and effort into the concept and creating it, and got what I wanted, so I think this is just a case of de gustibus non est disputandum. I felt it cleverly visually encapsulated the mood of the horror that Gene Wolfe meant to lurk underneath the harmless nearly-bucolic appearance of SD and enhanced the experience.
So I think it satisfies my 3 criteria: it is not low-meaning, was not low-effort, and adds something. But I don’t think this is a good place to discuss it, so I have added a more detailed discussion of that image’s process & meaning to my image slop blog post as an example of how I think I get good image samples.
EDIT: I would be curious about the disagrees. What, exactly, are you disagreeing with? Do you think I am lying about the creation process, the prompt, or the meaning? (I would point out that there was already a short version of this description in the alt text, and has been since I added it in the first place c. November 2023.) Do you disagree that the high concept reflects my SD interpretation? Or what?
If you asked me to guess what would be the ‘elegant’ counterpoint to ‘traveling with a carefully-curated of the very best prepper/minimalist/nomad/hiker set of gear which ensure a bare minimum of comfort’ was, I would probably say something like ‘traveling with nothing but cash/credit card/smartphone’. You have elegantly solved the universe of problems you encounter while traveling by choosing a single simple tool which can obtain nearly anything from the universe of solutions.
Remote seating has its own problems. On my last flight to SF, I almost missed my late-night flight out because I had (very unusually for me) found a more pleasant, quieter, empty gate to make a phone call on, out of eyeshot; and then my flight was delayed twice so the original boarding time flew past; and eventually I got so wrapped up in the call that I let the time slip until a vague nagging anxiety and had to wrench myself out and run in a panic to my actual gate—where fortunately there was still <10 minutes of boarding left. While it took at least 3 problems and I didn’t actually miss my flight in the end, it would’ve been bad because it was probably the last one out that night to SF, and it’s the first time I have ever come anywhere close to missing my flight while having actually been sitting at the gate hours before… So it was a memorable and alarming near-miss for me.
Indeed they are different mindsets. They are, however, both mindsets of concern.
I’ve never seen them express the slightest stress or mental tension. Zuiko of them is an old woman who’s hands hurt and are failing due to arthritis and she seemed more concerned with just listening to me than talking about her problems.
Sounds like a(nother) good example of the downsides of lacking pain/stress.
it mostly communicates that they’re a heavy user of the internet.
That true fact in no way contradicts dbohdan’s point. (“The import of an act lies not in what that act resembles on the surface, Mr. Potter, but in the states of mind which make that act more or less probable.”)
What am I missing? This seems like a big ol’ hunk of power, social capital, and utility, just sitting on the ground of central station!
This sounds like one of the riskiest and most thankless kind of (legal...?) websites which it is possible to run in exchange for not even a will-o-the-whisp of profit or payday.
After reading about Craigslist and OnlyFans and Facebook and Instagram and OKCupid and Backpage and Kuro5hin and MeFi and Tea and Mastodon-everything etc, I think I would rather slam a door on my private parts than sign up to try to create or moderate such a site for sub-minimum wage (at a total opportunity cost of likely millions of dollars, exclusive of psychic damage), and have to deal with people like, well, your following comment, and whatever regrettable decisions they inevitably will make.
For these reasons, I don’t really find the metrics used in the papers ad hoc, except to the extent that “award partial credit for answers that are close to correct” is ad hoc.
If they are not ad hoc, what have they successfully predicted in the two and a half years since they concocted their original reparameterizations like ‘BPE edit distance’ to ‘explain’ past emergence?
More links on this topic: https://gwern.net/note/competence
The Umeshism I always disagreed with the most, and agreed with even less after I actually missed some flights (mostly not my fault)...
I’ve wondered if Umesh was just much richer than I, secretly a Buddha, or if the changes in the US air travel system since Umesh was presumably doing most of his travel (pre-9/11, switch from heavily regulated/subsidized direct flights with wasteful regular departures to hub-and-spoke with connections, luxury to mass affordability downplaying of travel agents, and increasing fragility/delays/cascades of failures) means that we are comparing apples and oranges here.
(In particular, I’m thinking of accounts I’ve read of how a lot of flights used to be more ‘bus’-like, where you just hopped onto whatever plane happened to be there when you bothered to stroll in, and they departed 90% empty routinely. I saw an example of this in Hawaii on the local commuter airlines where my sister missed her flight to the next island but it was no big deal because there was another leaving >90% empty in an hour anyway so she just had a snack and chatted with us as we waited for a real flight out—the sort of flight that missing has consequences like ‘missing your connection in LA’ or ‘non-refundable ticket’. If all my flights were like that, and I wasn’t “missing” occasional flights, I would agree with that Umeshism!)
If only they had gone just a little bit further… https://en.wikipedia.org/wiki/Manicule
One reason is defensive, against lawsuits. When you license a book or ‘life rights’, you get a defense against someone saying ‘you stole my story’. (See for example Terminator getting successfully extorted by Harlan Ellison.) So they may actually go around and buy the movie rights to things that everyone knows is unrelated just to preemptively buy out any possible lawsuit grounds. While this may be ‘pennies from heaven’ for many writers, it has at least once had rather perverse consequences for one of my favorite writers, R. A. Lafferty, posthumously in creating a tragedy of the anticommons: https://ralafferty.tumblr.com/post/55382042501/49-the-six-fingers-of-time AFAIK, this also explains the bizarre practice of buying ‘movie rights’ for nonfiction reporting; for example, one of the Silk Road 1 movies was based on ‘licensing’ some articles from Wired. Or of ‘licensing’ public domain characters like Sherlock Holmes when you want to make a movie about Sherlock Holmes and are worried about an estate suing you over elements like ‘Holmes showing human emotions’ which are supposedly derived from still-copyrighted stories.
And yeah, in a thousand years a vampire could probably figure out how to protect themselves pretty well, so to write a story where the average guy wins, there must be a bit of stretch somewhere.
Interestingly, this is another point in which Bram Stoker’s Dracula is very well thought-out. Stoker is well aware that with his rules, Dracula ought to be invincible… But Dracula has the liability that he’s been stultified mentally by centuries of quasi-imprisonment, and so hasn’t yet understood or experiments with his powers.
He is slowly waking up, and doing so, and starting to understand that he can eg. move his coffins himself without hirelings, but only right as the protagonists hunt him down. It is only by hours or minutes do they manage to cut him off from each resource. With another day or two, Dracula would have realized he could, say, just bury a bunch of coffins deep underground in the dirt, and he would be immune from discovery or attack.
Really, the novel is shockingly rationalist, and that’s why I call it ‘the Vampire Singularity’. Dracula is undergoing a hard takeoff, as it were, which is just barely interrupted by the protagonists.
Minor correction here: ‘Hakuin’ could not have intended anything by the book, because he died in 1769 and it is his tradition being criticized. All the sources on The Sound of One Hand seem to attribute it to a (still unknown?) pseudonymous Japanese Buddhist author in the 1900s.
With stuff like rhyming, the issue isn’t tokenization, its that the information was never there in the first place. Like Read vs Bead. No amount of zooming will tell you they don’t rhyme. pronounciation is “extra stuff” tagged onto language.
Wrong. Spelling reflects pronunciation to a considerable degree. Even a language like English, which is regarded as quite pathological in terms of how well spelling of words reflects the pronunciation, still maps closely, which is why https://en.wikipedia.org/wiki/Spelling_pronunciation is such a striking and notable phenomena when it happens.
I think this is general and well defined. Like with images you could have it be just a literal zoom tool.
Shape perception and other properties like color are also global, not reducible to single pixels.
‘Repetition’ is certainly a drawback to the ChatGPT style: we have lost em dashes and tricolons for a generation. But it can’t in its own right explain the reaction to the SD image, because… ‘German Expressionist linocut’ just doesn’t describe a default, or even a common, output style of any image generative model ever. (That’s part of why I like to use ‘linocut’ as a keyword, and for better or worse, people who might reach for ‘German Expressionist’ these days typically reach for Corporate Memphis instead.)
It could however be a kneejerk reaction: “oh no, this is a generated image, therefore it is exhaustingly overused and boring [even if it isn’t actually]”.