One consistent tell is in hair. The most obvious manifestation is how long hair splits and merges back into itself (as at 0:27).
AI hair often doesn't follow proper "form" and looks like overcomplicated, random tufts that can clash and overlap. It might not be obvious on a first glance, but you can learn to tell the difference pretty quickly.
You say those things like they're a short step away, but that might not be how it works out.
For example, AI has made zero progress in the last few years in surpassing professionals at art or writing. Its prompt-following skill is much better, and sure, it can render hands and text now, but its artistic sensibility is completely stagnant.
I love this :) Thanks very much for making it, it's elegantly designed.
Since you asked for feedback: in terms of usability, I found the 'seek next' and 'seek previous' buttons confusing, since they're positioned left/right but control motion up/down, and even switch their direction based on loop. (This is because "forward" and "back" also change based on loop -- an indicator for that would help.) Adding navigation via mouse wheel would be perfect here too.
Sorry to ask for even more, but I'd personally love to see door opening / door closing sounds added (along with 'ドアが閉まります' and the alarm) to fully round out the soundscape.
Don't mean to be too picky! -- it's very enjoyable as is.
Also, studios with their own engine may release one or two fresh-feeling games, but across repeated releases, a custom engine is going to become strangling and repetitive way faster than any off-the-shelf option.
It's one of my favorite visual novels and is so underappreciated.
Maybe it's because (unlike others in the VN space) it totally eschews unusual settings, gimmicks, or flashy set pieces to sell itself... I only bought it because I liked the tidbits of story in Shenzhen and Infinifactory, by the same author. Every part of it is unbelievably strong though.
For me at least, the post is kinda confusing and feels a bit overwrought (“there exists a raven such that the vector of hours”?), and was hard to understand at first. Sadly, in the wonderful year of 2026, I can’t help myself wondering if it was all written by an LLM, prompted by “be mysterious” or similar — though I still wouldn’t bet on it.
The project is cool! It’s a simple visual graph layout system for making your own clock.
That’s so far from obvious. The most concerning possibilities for me — like kids not learning how to struggle or problem solve on their own — won’t be resolved for many years.
Your English writing seems to be quite excellent as well.
We have to draw the line somewhere with art, of course, and that usually comes down to a combination of what consumers value, and broader perceived cultural merit. Nobody really cares about the well-being of artists who specialize in making pretty paper airplanes, or drawing pictures using only the MSPaint pencil brush.
I think the audience, not the tools, deserve the most scrutiny here. Look around at this very thread, and all the people defending what the LLM wrote. Their feelings can't be argued with. But they make me feel sad and alienated, because I see a vast difference, so vast and so obvious, and they see none at all.
In the future, perhaps people will enjoy LLM work -- genuinely enjoy it -- as much as I've enjoyed Vonnegut or more. It may be the inevitable result of a broader cultural shift away from reading and writing. I guess with time, maybe we'll find out how valuable it actually is to have a strong command of one's language... I imagine, at least, people today are much better at other things to compensate.
There's no way. Just the first paragraph alone is enough to convince me; it's too well-written and melodious to be AI, with too much original thought:
Today, the demonic vice of the old is not that they are hard and demanding on the youth — instead they do not demand enough from us, and they cannot quite believe that we have not lived up to the little they have demanded. They think too well of our generation.
Without defending the quality of the rest of the essay, it's a great start. LLMs today could never match it.
I’m downvoting because it’s an unusual (and probably false) claim made with no evidence — particularly your clause after “as” needs a more substantive defense. Can you convince me a bit that you speak for the younger generation?
Yes, this isn’t a logical error at all. If you don’t have taste in one area — actually, it’s even worse, you’re not even aware of your own lack of taste — why would I trust your taste in another area?
Stephen's Sausage Roll is my favorite puzzle game. But more interestingly -- it's a near-universal opinion within puzzle communities that SSR is one of the all-time best. I've never heard of such a strong consensus in other subgenres of game.
Unlike other consensus "bests", it's relatively unknown to the public (which is understandable for many reasons). It's very likely that if you're a puzzle game devotee, you will fall in love with SSR; but at the same time, if you don't have experience with puzzle games, you'll very likely hate it.
As a result, I've always thought it's an interesting window into how we value "taste" and "mastery", how too much mastery can actually distance us from one another, and what meaning there is in designing games for an ideal world shaped around ourselves, versus the world we actually live in.
It's well-known that puzzle games sell badly on Steam, and I think part of that is that difficulty and struggle is an acquired taste. Most try to paper over that gap with nice soundtracks and graphics, "hooky" mechanics, and narrative. SSR is so interesting because it contrasts so violently: it's ascetic, has no obvious hook, and offers nothing but difficulty and struggle, and the best feeling in the world if you decide to push through it anyway.
It seems like the author is studying “disorder” in an array by restricting themselves to looking only at local differences. They make up a lot of reducer functions, many of which relate to each other. But I don’t see the usefulness in playing with them — as none of the individual functions are ever useful. The author even says in their article about Amp that it doesn’t really work.
Basically all the functions only observe local behavior, but to measure “disorder” in a sensible way you need to take into account global relationships; the disorder of 2-3-4-5-6-5 is different than that of 2-3-4-5-6-1. There doesn’t seem to be a proposed way around this, but maybe I missed something.
Hoping to release in 2027 :)