As written, at least for us, this already IS a serious issue. I think there is simply not enough German-language AI projects yet, and so we Germans prefer to discuss real pressing matters such as the 'Pendlerpauschale'… o.O
Over the last six months, me and friend tried to automate German comedy. It’s a long read, but I thought it deserved a proper write-up because, instead of just speculating about AI in business, we actually built it… questions very welcome!
"derweil is an interactive video installation correlating time, space and big data to provide tailor-made instructions on how to get lost.
Materials: Google Directions and Streetview APIs, JavaScript, NW.js, cables.gl, Involt, Arduino IDE, computers, thermal paper, plastic, metals, wood."
(I doubt I could still run this today, though. I used some kind of 'hack' to bypass Google Streetview API limitations and I'm pretty sure they fixed this ages ago… ;D)
Mpv is great! And also the only player I found to be externally controlable, e.g. through a hardware jog/shuttle or an Arduino. This might be outdated (from 2018), but for reference (me talking to myself ;): https://softwarerecs.stackexchange.com/questions/53208/video...
Cables is absolutely fantastic. I used it personally for an art project, as well as was involved with a commerical AR experience which used cables to run elaborate, fully interactive 3d scenes in a normal browser, on mobile. As with other node based languages (e.g. vvvv, max/msp), you edit your code while it's running, so you directly see what you're doing without constantly switching interfaces. And in the end it generates a js file you can just embed in an iframe. Honestly, no idea why this is not more widely used, huge fan!
…well, if it runs. Because all these predictive-auto-whatever features also break things: Eg., I have a bug in Apple Mail [1], which basically breaks "entering text into a computer using a keyboard" – a problem I would have thought was solved some 70+ years ago, but alas, here we are…
I read almost all of this, and am kind of torn on this ressource: On one hand this is a great introduction to ComfyUI, including concepts and possibilities, and it really IS helpful for getting started. Thanks to the authors for putting in the work!
That said, some of the information, especially regarding the more technical parts, is somewhat misleading (debatable, simplification is hard), and at times outright false (e.g. "This method is called LoRA (Loopy Recurrent Attention)"… what?), so don't make it your only source of info.
Eg. I found this also helpful: https://blenderneko.github.io/ComfyUI-docs/
For the second part of your question regarding style and postprocessing: In terms of style, what really helped was to use "--style raw" and also to give it very specific locations like "…in a contemporary art museum". I had a whole bunch of these which where randomly added to the prompt. In terms of postprocessing, I upscaled the images with Stable Diffusion locally (using imaginAIry) and then added some grain, but that's it.
Thank you! I did not use the concepts directly as prompts, but did ask the concept-model to also output a visual description (which was only used internally). This description I then piped through GPT4 (which was much better at this then GPT3), then added some modifiers, and only then sent it to Midjourney. You will often note though, that language comprehension is not really on point – like often it does show the right elements in the image, but not necessarily in the same relationship to another as they are described in the concept...
I had an art exhibition last week, where I tried to replace myself as an artist with a Henry Ford-style production line (including a moving conveyor belt), run by different AI models to produce my art. The system looks for inspiration in my emails, my bookshelf, surfs reddit and also HN, and then creates pretty fancy installation concepts plus the matching installation views. The 125 artworks it produced within a day and a half can all be checked out here: https://maschinenzeitmaschine.de/p_doppelgaenger_gallery/
Under the hood, it's Python (Flask), mostly using GPT-3 & 4 plus Midjourney (with fallbacks to DALL-E and local Stable Diffusion as using Midjourney programmatically is pretty fragile). I tried fine-tuning models, but as I did not have enough data, I got better results just chaining complex, modular prompts. The conveyor belt was made out of spare parts for treadmills… feel free to ask any questions!
I made an interactive art installation on this question once: A black box with a knob where you could adjust how much time you have, then it would offer you (Google streetview) panoramas of the locations it found within your distance, and in the end even print a paper slip with your travel itinerary to take with you.
The installation used realtime data (Google directions API): I somehow figured out, that if I would run this from a local machine and reset the browser frequently, Google would let me do this even without an API key… they certainly sensed something was awry and I did get API warnings and captchas because of 'suspicious traffic on my network', but they were nice enough not to block me completely. I strongly doubt this would still work though, this was in 2017.
I'll probably get downvoted into oblivion for mentioning blockchain tech, but this might at least help, maybe not in its current form but... I did not follow that project, but there do exist some concepts in this direction, e.g. this: https://techcrunch.com/2021/01/12/numbers-protocols-blockcha...
Yes it was, but is it automatically bad because of this? It would be if it would be unhelpful, low-quality or if it would contain deceitful content pushing their product – but I do not see any of this.
The article is an overview aimed at beginners, but as such actually pretty good and helpful. It does not seem to overtly promote their product. Classifying this as spam / SEO only because one is not the target group is not fair.
I'm not the type of person who would maintain such a catalogue, but thinking about the scenario: I'll take 15 min tomorrow to run through my flat and just photograph everything. That way, in case disaster strikes (knock on wood it never does!), I do have some proof and only have to do the tedious cataloguing when it's actually needed… so, thanks i guess!