101–110 of 279 posts
Earlier quoted context omitted.
This probably isn't backed by an LLM but instead some kind of geometric shape model.
How do you explain a horse 2 legs become 4 legs when rotated assuming they only drew 2 legs on the side view
They could be using transformers, sure. But plenty of transformers-based models are not LLMs.
Earlier quoted context omitted.
This feels extremely ungenerous to the Big Tech companies. What's wrong with trying out 100 different AI features across your product suite, and then seeing which ones "stick"? You figure out the 10 that users find really valuable, another 10 that will be super-valuable with improvement, and eventually drop the other 80. Especially when if Microsoft tries something and Google doesn't, that suddenly gives Microsoft a…
> What's wrong with trying out 100 different AI features across your product suite, and then seeing which ones "stick"? Even the biggest tech companies have limited engineering bandwidth to allocate to projects. What's wrong with those 100 experiments is the opportunity cost: they suck all the oxygen out of the room and could be shifting the company's focus away from fixing real user problems. There are many other pr…
If you look at all the recent Google Docs features rolled out, only a small minority are AI-related:
https://workspaceupdates.googleblog.com/search/label/Google%...
There are a few relating to Gemini in additional languages and supporting additional document types, but the vast majority is non-AI.
Seems like the companies are presumably sorting on ROI just fine. But, of course, AI is expected to have a large return, so it's in there too.
The source (Adobe MAX) 'demoes' full range of incredible scenarios.. https://www.youtube.com/watch?v=gfct0aH2COw
Not bad, but what's up with the audience? Is there an Adobe cult or something?
Earlier quoted context omitted.
In some way, having followed the open source image generation scene for a while, it feels a little bit like it's opposite? Most of the ai image generation stuff I've seen from adobe feels late to party in terms of what you can do with open source tools. Where they do compete however is with tight integration, and I guess that's what matters the most to users in the end. There are plugins for gimp that let you do imag…
I think your last paragraph sums it up pretty nicely: users need a good UX to get to these tools. So I would love if GIMP started shipping these awesome plugins by default to pick up the pace!
Incredible, but a shame you'll have to use Adobe to get it.
They were double charging me for photoshop for two years. I caught them and it took 60 minutes on the phone to get them to do something about it.
They have an entire cancellation department. (!)
I find that Adobe is really pulling away from open source software with all this AI stuff. A few years ago it could be argued that GIMP, Inkscape, and Darktable could do almost everything that Photoshop, Illustrator, and Lightroom could, albeit with a jankier user interface. But now none of the open source software can compete with AI generative fill, AI denoising, and now AI rotation.
Ex - I can absolutely get exactly this same rotation feature using open toolchains, they just haven't been nicely consolidated into a pretty package yet.
So to recreate the same thing adobe is doing here I currently have to:
1. Use the 3d-pack in comfy-ui to get stack orbit camera poses for my char (see: https://github.com/MrForExample/ComfyUI-3D-Pack scroll down to stack orbit in the readme)
2. Import those images back into the open source tool manually.
Is it as convenient? Nope - it requires a lot more setup and knowledge.
Is it hard to imagine this getting implemented in open source? Also nope. It's going to happen, it just won't be quite as quick.
Earlier quoted context omitted.
> What's wrong with trying out 100 different AI features across your product suite, and then seeing which ones "stick"? For users ? Almost everything is wrong with that. There are no users looking for wild churn in their user interface, no users crossing their fingers that the feature that stuck for them gets pruned because it didn't hit adoption targets overall, no users hoping for popups and nags interrupting their…
Think about the other side though -- if the tool you've learned and rely on goes out of business because they didn't innovate fast enough, it's a whole lot worse for you now that you have to learn an entirely new tool. And I haven't seen any "wild churn" at all -- like I said in another comment, a few informative popups and a magic wand icon in a toolbar? It's not exactly high on the list of disruptions. I can still…
I spent so many hours trying to do rotations with a pirated copy of Flash as a kid, and I never really got the hang of it, and it always bothered me how deceptively hard rotation was; when I would show my parents my work, they would do their very best to try and act excited but I could tell that they weren't really impressed with the effort because it doesn't seem that hard, at least to a lot of people. This makes me…
Earlier quoted context omitted.
Force-feeding 100s of different AI features (90% of which are useless at best) to users is what's wrong with the approach.
Why? It's not "force-feeding". You usually get a little popup highlighting the new feature that you close and never see again. It's not that hard to ignore a new "magic wand" button in the toolbar or something. I personally hardly use any of the features, but neither do I feel "force-fed" in the slightest. Aside from the introductory popups (which are interesting), they don't get in my way at all.