FinishingDutch @ FinishingDutch @lemmy.world

Posts

1

Comments

545

Joined

2 yr. ago

They actually neutered their AI on thursday, after this whole thing blew up.

https://abcnews.go.com/Business/wireStory/google-suspends-gemini-chatbots-ability-generate-pictures-people-107446867

So right now, everyone's fucked because Google decided to make a complete mess of this.

Well, the tech is of course still young. And there's a distinct difference between:

A) User error: a prompt that isn't as good as it can be, with the user understanding for example the 'order of operations' that the AI model likes to work in.

B) The tech flubbing things because it's new and constantly in development

C) The owners behind the tech injecting their own modifiers into the AI model in order to get a more diverse result.

For example, in this case I understand the issue: the original prompt was 'image of an American Founding Father riding a dinosaur, while eating a cheeseburger, in Paris.' Doing it in one long sentence with several comma's makes it harder for the AI to pin down the 'main theme' from my experience. Basically, it first thinks 'George on a dinosaur' with the burger and Paris as afterthoughts. But if you change the prompt around a bit to 'An American Founding Father is eating a cheeseburger. He is riding on a dinosaur. In the background of the image, we see Paris, France.', you end up with the correct result:

Basically the same input, but by simply swapping around the wording it got the correct result. Other 'inaccuracies' are of course to be expected, since I didn't really specify anything for the AI to go of. I didn't give it a timeframe for one, so it wouldn't 'know' not to have the Eiffel Tower and a modern handgun in it. Or that that flag would be completely wrong.

The problem is with C) where you simply have no say in the modifiers that they inject into any prompt you send. Especially when the companies state that they are doing it on purpose so the AI will offer a more diverse result in general. You can write the best, most descriptive prompt and there will still be an unexpected outcome if it injects their modifiers in the right place of your prompt. That's the issue.

I’ve never even heard of that, so I’m definitely going to check that out :D I’d much prefer running my own stuff rather than sending my prompts to god knows where. Big tech already knows way yoo much about us anyway.

Here’s one that was made, just for you, with specifically a VEGAN cheeseburger in the prompt :D

Doesn’t look too bad to me. I love a fair bit of crispy lettuce on a burger. Doing it like that at least spreads it out a bit, rather than having a big chunk of lettuce.

Still, it that was my burger… I’d add another patty and extra cheese.

There’s also good fun in derailing a thread :D

The other day I commented on a post that used an ‘internet pirate’ stock image. Mentioned how I’d seen a guy with an eyepatch on the street that day and how rare that was.

Someone ended up sharing a neat little story about getting poked in the eye with a sword and how a kid called him a pirate. That all added fuck-all to the topic at hand, but it sure was collectively entertaining.

Most people are guilty of that to some extent. You want to make a good response, but end up overthinking it and moving on. Which is a shame. God only knows how many awesome comments end up in the digital wastebasket.

Besides, what’s the worst that can happen? Who really gives a fuck about downvotes anyway.

Exactly. This thread is a perfect example. There’s literally no real topic to discuss, and yet people are talking. And that’s a great thing to encourage if we want to grow this platform 👍

Well, comment anyway. You never know what’ll happen.

And even if you have a boring, vanilla opinion on that topic, post it anyway. Because it’ll lower the bar for others to comment as well. (As this entire thread demonstrates)

Reddit for sure is toxic. Generally, it’s much easier to be toxic in a large, anonymous group with an endless amount of subreddits to retreat to. Here, it’s maybe 10-20 people talking, so there’s not much room to hide, as it were. You keep running into the same faces, so it’s a bit more important to stay polite.

Funnily enough, he’s not eating one in the other three images either. He’s holding an M16 in one, with the dinosaur partially as a hamburger (?). In the other two he’s merely holding the burger.

I assume if I change the word order around a bit, I could get him to enjoy that burger :D

Lemmy could definitely use a bit more comment activity on a lot of posts.

I think it’s because nobody really wants to be the first to comment and offer an opinion that might end up going against the grain when a thread develops. There’s no ‘reading the room’ as it were.

I’m doing my part by commenting on threads. Like this one.

Because sometimes you want an image of George Washington, riding a dinosaur, while eating a cheeseburger, in Paris.

Which you actually can’t do on Bing anyway, since it ‘content warning’ stops you from generating anything with George Washington…

Ask it for a Founding Father though, it’ll even hand him a gat!

Speaking for myself, it’s definitely not the lack of detail in the prompts. I’m a professional writer with an excellent vocabulary. I frequently run out of room with the prompts on Bing, because I like to paint a vivid picture.

The problems arise when you use words that it either flags as problematic, misinterprets anyway or if it just injects its own modifiers. For example, I’ve had prompts with ‘black haired’ rejected on Bing, because… god knows why. Maybe it didn’t like what it generated as it was problematic. But if I use ‘raven-haired’ I get a good result.

I don’t mind tweaking prompts to get a good result. That’s part of the fun. But when it just tells you ‘NO’ without explanation, that’s annoying. I’d much prefer an AI with no censorship. At least that way I know a poor result is due to a poor prompt.

I’ll look into it for sure. I tried Automatic1111 last year with SD, bunch of add-on stuff… it was finicky and didn’t get me quite what I was looking for.

Thanks for the tip!

Honestly, this sort of thing is what’s killing any sort of enjoyment and progress of these platforms. Between the INCREDIBLY harsh censorship that they apply and injecting their own spin on things like this, it’s nigh on impossible to get a good result these days.

I want the tool to just do its fucking job. And if I specifically ask for a thing, just give me that. I don’t mind it injecting a bit of diversity in say, a crowd scene - but it’s also doing it in places where it’s simply not appropriate and not what I asked for.

It’s even more annoying that you can’t even PAY to get rid of these restrictions and filters. I’d gladly pay to use one if it didn’t censor any prompt to death…

I absolutely love that story!

A tall angry dude dressed all in black leather with an eyepatch definitely would give me Snake Plissken vibes; either that or someone from an anime :D

Please do wear one if you want to! That guy made my day; he looked like a badass.

It’s such a rare thing to see, you’ll probably get a few jokes. But also some really fun interactions I imagine. If nothing else, you’ll give some kids a fun story to tell about that pirate they met :D

You know, I saw a guy the other day with an honest to goodness eyepatch. First time I’ve ever seen one in the wild.

Guy looked like older Snake from Metal Gear Solid, but clearly wasn’t cosplaying or anything like that. I assume most people with one eye just get a fake one here. If only to avoid all the pirate jokes that he probably gets.

Well, all I can reasonably do is put it in the correct bin for proper recycling, as they tell you to do. I would never litter; it’s antisocial as fuck.

Once it’s collected, I have zero control over it. I assume Dutch garbage companies don’t dump my trash in the ocean, but I honestly haven’t bothered to check if they export it to somewhere else or anything like that.