Lightning_Shade

I say what I think. Politically wrong-wing.

Ask not what art can do for the world, ask what art can do for you.

X (former employee), Y (current), Z (future)

[Lightning_Shade @LightningShade0]

Suffocate the voices of preference falsification. Strangle the internal liar. Stab the coward within into the miserable void it pretends is "your heart".

Like and dislike art without shame. Know that pleasure is Good and Just. Let no rules stand in your way.

[Lightning_Shade @LightningShade0]

@kukuruyo Also, the one intentional change that *is* there is interestingly subtle... https://t.co/JcIS2QkqyU

[Lightning_Shade @LightningShade0]

I just want to highlight the one intentional change I see here -- "please make a more angular tsurime-ish eye shape" is a *very* subtle detail. Whoever was doing it nailed the assignment, for sure.

[Lightning_Shade @LightningShade0]

I just want to highlight the one intentional change I see here -- "please make a more angular tsurime-ish eye shape" is a *very* subtle detail. Whoever was doing it nailed the assignment, for sure.

[Aurondarklord wants Kairi Sane back! @Aurondarklord]

Claire and the woman they face scanned for her. Very close to exact. There is no technological problem that stops western devs from doing this, uglification is on purpose.

[Lightning_Shade @LightningShade0]

@Aurondarklord Considering how important the "tsurime eyes/tareme eyes" distinction is in anime and manga, I feel very confident that pushing the final model into "as close to full tsurime as possible while maintaining realism" is intentional.

Very interesting comparison, thanks. /2

[Lightning_Shade @LightningShade0]

@Aurondarklord Not only is it very close to exact, but the most important change that *is* there is *subtle* in a way we keep hearing can't be done.

The final character has a more angular eye shape (and non-rounded eyebrows), which HAS to be deliberate. /1

[Lightning_Shade @LightningShade0]

@tracewoodgrains On the other hand, humans didn't *stop* playing chess just because Stockfish exists. The human spirit is resilient.

[🎭 @deepfates]

The student finished reciting his koan and looked eagerly at the master. The master looked annoyed.

"I don't do koans, as a rule," said the master.

The student said "What do you mean?"

The master said, "Think about it."

The student thought about it for a long time. He considered it from all angles, looking for paradoxes, subtleties, allusions. He began to grow distressed, as it didn't seem enlightening at all.

Eventually he realized that the master just fucking hated koans. Then he was enlightened.

[Lightning_Shade @LightningShade0]

@feelsdesperate Read it in Alex Jones' voice, it becomes funnier.

[░ perfectloop ░ @PERFECTL00P]

𝙲𝚘𝚙𝚢𝚒𝚗𝚐... 📁📰📂🟦

[Lightning_Shade @LightningShade0]

@DanMcGarryArt Did humans stop playing chess because Stockfish exists?

[Lightning_Shade @LightningShade0]

@neogeo8man That's not unique to Star Fox 64, is it? I think speedruns in general have way outstripped scoring runs in terms of popularity, gaming-wide.

(Which is not great for games where scoring is the more interesting competition, I guess.)

[Lightning_Shade @LightningShade0]

@tjbierschbach Anyway... we do have a lot of "do not ever enter anything sensitive into ChatGPT" stuff in trainings, but they don't actually prevent us from using it when it's *not* sensitive data. I do know many companies just block it entirely (understanadable), but not where I work.

[Lightning_Shade @LightningShade0]

@tjbierschbach I tried it a fair amount of time ago and back then I think it still used GPT-4, and it really couldn't do even those small things reliably well. Time sure flies.

[Lightning_Shade @LightningShade0]

@tjbierschbach If you're careful about what you ask, you can get answers to small specific bits of functionality that involve nothing "identifiable" at all. (Could I do that myself? Absolutely, but it'd take hours because I'm not very familiar with Powershell syntax.)

[Lightning_Shade @LightningShade0]

@tjbierschbach "How do I make a Powershell script that runs this test command with these options, saves the output to a file, corrects encoding to utf-8 and strips ANSI control codes" exposes no client data. (Assuming "options" don't have them, of course.)

[Lightning_Shade @LightningShade0]

@Aurondarklord @grok Huh, looks like we are *both* right.

1) The feminized "Hawkeye Initative" pose really isn't a good match (as you said)
2) But even then, the shading *still* makes it easier to treat it as a serious character drawing (as I suspected).

(I do like Grok's take on the background.)

[Lightning_Shade @LightningShade0]

@tjbierschbach Agentic Copilot-like stuff is wildly more expensive now and it's out of line (true), but on the flip side, ChatGPT's free model has become a hell of a lot better. For "simple stuff in unfamiliar syntax" (e.g. a Powershell script), it can actually hard carry now, unlike before.

[Lightning_Shade @LightningShade0]

@Aurondarklord Too bad the artist was lazy and didn't even try, so neither of us have a way to *check* this, eh?

[Lightning_Shade @LightningShade0]

@Aurondarklord Personally, my mind *still* jumps to "where's the shading", and I think it'd change perception somewhat... but "this isn't how women depict men in their own fanservice material" is true either way AND more important, so I think at this point I'm just splitting hairs.

[Lightning_Shade @LightningShade0]

@Aurondarklord My hot take is actually that even those exact poses would usually look 200% times less weird if the drawing was shaded properly in the original's style. It usually isn't, because it's a shitpost strawman and shading takes effort.