HumanAIGC
Fraud in the AI industry, Vaporware aka "Proof of concept, or GTFO"
Too much of this going on.
It's all too easy to fake results and attain funding - this needs to stop!
https://github.com/magic-research/magic-animate for a working alternative
This isn't faked. You can still see artifacting, unnatural things going on and theyre all perfectly in sync. Please take a look at the results
Too much of this going on.
It's all too easy to fake results and attain funding - this needs to stop!
https://github.com/magic-research/magic-animate for a working alternative
Before insulting people, their work, and making a fool of yourself, you should research the individuals who created this document.
If they released it and it works as demoed in the video... everyone will return here and have a good laugh. (myself included)
And I will apologise, sincerely.
I'm happy to be proven wrong - but do please understand that this is a systemic issue within the growing space.
Whether it's vulnerability exploitation, blockchain projects or even LK99 claims - there has to be pushback from the community for proof of concept. This cannot be the new norm.
I'm happy to be proven wrong - but do please understand that this is a systemic issue within the growing space.
Whether it's vulnerability exploitation, blockchain projects or even LK99 claims - there has to be pushback from the community for proof of concept. This cannot be the new norm.
Literally look at the results. They are not perfect in any way, you can see obvious arrifacting when you look at these results. Please do your research before going "oh. Fake" when you see the first few results, not even looking for flaws.
We can indeed see artifacting, this is true. How the videos were produced, nobody can tell.
As an aside, please remember to put the Python version number in the Readme when you do launch - seems to be another issue within the community.
@chrisbward what exactly are examples of fraud in academic community researching AI topics? And do you really think that Alibaba researches publish this paper to attract funding?
https://github.com/guoqincode/AnimateAnyone-unofficial https://github.com/MasahideOkada/animate-anyone
🔥 https://github.com/dreamoving/dreamoving-project
This isn't faked. You can still see artifacting, unnatural things going on and theyre all perfectly in sync. Please take a look at the results
I dont think he means like, its real ppl dancing and they faking it. I think he means, its pretty easy to do these sorts of things with already available tech, and make out like you have created something 'new' in order to get funding etc
Github turning into a place for teams to advertise, and at any point now, its gonna turn super fake, super fast.
This isn't faked. You can still see artifacting, unnatural things going on and theyre all perfectly in sync. Please take a look at the results
I dont think he means like, its real ppl dancing and they faking it. I think he means, its pretty easy to do these sorts of things with already available tech, and make out like you have created something 'new' in order to get funding etc
Github turning into a place for teams to advertise, and at any point now, its gonna turn super fake, super fast.
Precisely this. I am posting the exact same sentiment on all projects that follow this pattern. It's not to target this one per-se.
It's a larger issue with where the community and industry is headed!
Though honestly, I'm not sure AA is the best example of that anyway. It isn't like there was a big marketing campaign that I know of. I think there was a few people that watch papers coming out who profiled it first and then all the AI influencers on twitter decided to run with it. I don't think Alibaba even made a tweet on it themselves. I wouldn't be surprised if the researchers felt a bit sideswiped by all the sudden attention. Other similar projects like DisCo and DreamPose never got that same attention that I know of.
@chrisbward So, that means you're harassing multiple research teams with baseless accusations of misconduct? I really don't understand why people choose to come and be angry at a team that haven't invited anyone, doesn't sell anything, don't need public funding, just sharing their research. No one advertising anything on github, the repo isn't even for users, it's for fellow researchers.
Expectations are becoming more unrealistic every day. Not only it should be product quality from the day one, also completely free, open-source and there still be accusations of fraud or at least of being part of the trend "that there soon be only fraud".
Chillout and let the people work.
@chrisbward So, that means you're harassing multiple research teams with baseless accusations of misconduct? I really don't understand why people choose to come and be angry at a team that haven't invited anyone, doesn't sell anything, don't need public funding, just sharing their research. No one advertising anything on github, the repo isn't even for users, it's for fellow researchers.
This repo hit popular press. So random people came to see, not knowing that the standard in these things is to publish first, and then put up code (if code is delivered at all) a week to a couple of months later. The good stuff is in the math in the paper, of course. But script kiddies want to get something to run that they don't understand.
This project should not have opened issues until the code dropped. Might be a new thing going forward. @HumanAIGC should definitely close this one.
@sroussey yeah, maybe. Quick look at Chris Linkedin reveals that he is some kind of consultant in generative content area, so he probably should know better then demand freebies from industry researchers while basically building business on their work and accusing "industry" of fraud.
@sroussey yeah, maybe. Quick look at Chris Linkedin reveals that he is some kind of consultant in generative content area, so he probably should know better then demand freebies from industry researchers while basically building business on their work and accusing "industry" of fraud.
Actually a dev of 23 years, I have a couple of side projects, I contribute a lot to open-source projects. I'm not demanding any "freebies".
I have some experience in exploit development and this shit doesn't fly in those circles.
I'm closing the ticket, but I wouldn't hold my breath.
These could be generated by an i stagram filter with face tracking, which seems more likely.
Alibaba doesn't inspire confidence by showing examples with no explanation
How do you mean no explanation. You know there is a whole detailed research paper devoted to this right? https://arxiv.org/abs/2311.17117
And people have already implemented open source versions based on the paper. Not as good as the original, but still showing that it works in general.
These could be generated by an i stagram filter with face tracking, which seems more likely. Alibaba doesn't inspire confidence by showing examples with no explanation
How do you mean no explanation. You know there is a whole detailed research paper devoted to this right? https://arxiv.org/abs/2311.17117
And people have already implemented open source versions based on the paper. Not as good as the original, but still showing that it works in general.
Dont worry, these people don't actually know what a research paper is. No wonder they didn't find one
6 months, no code - clowns
You're 100% correct. You called it at the start, and I mistakenly doubted you.
