I can’t say I understand what people get out of playing with LLMs, but if they view writing code or drawing pictures like I view using a calculator then that explains a lot of the disconnect between the two camps.
I would 100% let some model draw me up a comic to shitpost without a second thought, and it by no means should be construed to mean that I wish I had chosen art for a career or a major or even a hobby.
Interestinger and interestinger. So you wish you could create the art (or pictures, whatever term you want to use), but the way you do it doesn’t matter, so you just use the LLM because that’s the easiest way to do it?
Would you normally try to soup something up in paint, or just do something else? I mean like before you could generate it with natural language.
I don’t personally use llms anymore, but when I did, it was for use cases where either I wanted a thing to be reasonably easy/fast or I would do without (a character sketch for DnD) or for visual representations that quick google searches weren’t turning up.
I watched a video where someone sent AI images to dress makers so they could make dresses. If I wanted a very specific type of dress, and I can’t draw this seems like a great way to communicate my desires and intentions.
I realize I’m a little bit of a liar. I will sometimes use Be My AI to help alt-text/caption pictures that happen to have a lot going on. So they get a list of snacks instead of something like “table with various snacks”.
So just FYI the rule of thumb with alt text that I’ve heard is: imagine you’re describing the picture you’re looking at to someone you’re talking on the phone with.
So that usually means keeping it brief and precise to only the aspects which are relevant to the context it’s being posted in.
Sorry if you’ve heard this already, but yet helped me a lot.
So if I was making a website or posting to Facebook it’s probably table of snacks. But since it’s people whose preference I know and who were there I’ll go all out and ask AI to give all details
Doubt.
You spend so much time using your clock and calculator on your phone you clearly want to be an engineer but don’t have the skills.
If that were true, there would be people saying they are engineers for doing this.
There aren’t.
But there’s people saying they’re artists and musicians for making slop…
That’s actually a really illuminating analogy.
I can’t say I understand what people get out of playing with LLMs, but if they view writing code or drawing pictures like I view using a calculator then that explains a lot of the disconnect between the two camps.
I would 100% let some model draw me up a comic to shitpost without a second thought, and it by no means should be construed to mean that I wish I had chosen art for a career or a major or even a hobby.
Interestinger and interestinger. So you wish you could create the art (or pictures, whatever term you want to use), but the way you do it doesn’t matter, so you just use the LLM because that’s the easiest way to do it?
Would you normally try to soup something up in paint, or just do something else? I mean like before you could generate it with natural language.
I don’t personally use llms anymore, but when I did, it was for use cases where either I wanted a thing to be reasonably easy/fast or I would do without (a character sketch for DnD) or for visual representations that quick google searches weren’t turning up.
I watched a video where someone sent AI images to dress makers so they could make dresses. If I wanted a very specific type of dress, and I can’t draw this seems like a great way to communicate my desires and intentions.
https://youtu.be/LeGcbc87JxU
I realize I’m a little bit of a liar. I will sometimes use Be My AI to help alt-text/caption pictures that happen to have a lot going on. So they get a list of snacks instead of something like “table with various snacks”.
So just FYI the rule of thumb with alt text that I’ve heard is: imagine you’re describing the picture you’re looking at to someone you’re talking on the phone with.
So that usually means keeping it brief and precise to only the aspects which are relevant to the context it’s being posted in.
Sorry if you’ve heard this already, but yet helped me a lot.
I upvote you because yes, but also I have a few light independent (non sighted) coworkers who like all the details.
So if I was making a website or posting to Facebook it’s probably table of snacks. But since it’s people whose preference I know and who were there I’ll go all out and ask AI to give all details
This is basically just taking a shit on the sidewalk and leaving it there.
Ok? So is Mimi comics and pizzacake if you ask me.