this post was submitted on 19 Nov 2023
497 points (86.9% liked)
Technology
59422 readers
2854 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
One of my jobs involved updating blogs for small businesses. I had a Shutterstock subscription for the images that goes along with these blog posts. For this task, I think AI generated images work a lot better than stock photography.
There's some recruitment company advertising jobs on LinkedIn. All the pictures are clearly AI generated and they're terrifying. Uncanny Valley freaks grinning at you from your screen.
you can already api into chatgpt and dall-e 3 as one cohesive service, and make a system in an afternoon's work that reads the article, decides on a thumbnail, and automatically generates one. the whole thing costs like 8 cents per article.
You could also self host Stable Diffusion to save some money
Edit: Or is it free to use with ChatGPT Premium (or whatever it's called)? Then that would actually be cheaper
no, the premium stuff doesn't give you api access. which is total bs, but yeah, it's only for that grey interface. (i'm also quite salty that the playground has no easy to access image inputs but that's beside the point)
you're completely right about self-hosting sd, it's just a matter of prompting. sd workflows tend to get a little more experimental but i guess you could still make chatgpt write a few prompts that are close to correct and just manually rerun if an image failed
Maybe ChatGPT knows how to write SD prompts at this point. You could try just telling it to generate a prompt specifically for SD.
yup, and when in doubt, just do some few-shot prompting
I still need to set up my computer to do the AI generated images.
My office is using AI to do this years xmas card. It's pretty cool.