I like the articles you and yog post currently. I browse Local, not Subscribed, and they often appear on my timeline. With Lemmygrad growing and more posts being made per day it’s natural that not everything will get seen equally.

Maybe people might be interested in more practical uses of AI, I know at least I am. A lot of stuff we already use is neural networks (what I mean by AI personally) like image upscalers, automated image coloring, speech-to-text (like Siri, but not huawei’s voice assistant and it shows and they should really do something about this because it sucks!), moderation tools so that humans don’t have to look at CSAM (very important), translation work is also getting very good with neural networks for all sorts of usecases but you don’t have to rely on google translate anymore when that tool had a 60% accuracy at best.

I’m just spitballing but I was just seeing a new tool that automates retopology for 3D artists, which is when you simplify a 3D model you sculpted to drastically lower its polygon count and make it easier to animate. And I was thinking there’s a lot of places in my own formal work where I could free up effort if I could have similar tools, like photoshop has genAI (if you use cloud which uh, let’s just say not everyone lets Adobe through their firewall…) but some filters and effects such as selecting a subject still rely on old algorithms and don’t work very well. Selecting hair especially has been refined over the years but it’s still a mess to perfectly isolate a subject, it takes a lot of time and the results vary. Being able to click a button and have a perfect selection would be dope, and I’m sure neural networks can take care of this.

edit - I also have some ideas for posts about AI, hopefully people won’t think I’m trying to troll by posting them (if I do end up making them)