lobsters score: 1 · tags: law
As a paying subscriber (which I will not renew) and someone who have had bad experience with ente before, this doesn’t surprises me, I was already looking for alternatives when I found Claude commits in their code base.
But I am more angry because Fairscan is one of hard to find app where author has really put genuine effort. I have been using it since its Inception and been reading author’s blog too, so I am biased.
Also I don’t know if this would get my subscription canceled or but this was my mail to ente in 2025 (mind the grammatical mistakes) and everything in this mail still stand, nothing has changed.
-
app size is 345mb on f-droid
-
they promote discord over their webpage and I don’t know about the matrix situation because I left the channel
The email:
On June 20, 2025 3:26:00 PM UTC, Yumm <redacted email> wrote: Hello ente, I am a paid user of your photo service, and I love it. But I have several concerns and requets regarding the serivce - As a service who priortise privacy and Open Source, having discord as your main exchange medium is concerning, I have sent my requets in the matrix channel quite a few times but no one from the team replied to that. Eventually I found out matrix channel is not looked after by the team and that saddening, As a paid user, I feel neglected even if I was a non-paid user I would want my problems be heard by the team. - The app size is over 350mb and that's also an issue on lower end devices and such. You should work on reducing the app size as a priority. Also you could app provide apps for diferent Architectures, that would reduce the size to half rather than one universal apk. I hope my requests will be heard this time.The reply:
Hi, thanks for the feedback! We've shared your concerns with our dev team and will be discussing them internally. <stuff related to jpeg-xl support i had requested in another email> Thanks again for reaching out and helping us improve! Best, Jay
score 1 · canonical
I too am a paying ente customer (200Gb plan) and am not happy about the ramping up of the inclusion of LLM generated code. They also seem to be spinning up new products every other month when the reason I chose them was solely photos. The experience of which still has plenty of room for improvement before diluting their focus with additional products. The most recent of which I’m dismayed to discover is “Ensu, our first step toward a private, personal LLM app that runs on your device and grows with you over time.”
They seem way too LLM brained for me. Zeitkapsl looks interesting, I’ll need to evaluate it further. One positive sign is they host their code on Codeberg, which doesn’t permit largely LLM generated projects. How has your experience been with it?
Edit: Not off to a great start. I tried building the desktop app as I’m on an uncommon distro and the Makefile references a target that seems to have been deleted 25 Sep: Makefile tries to run copy-models in
mlbut that Makefile was deleted in a commit from 25 Sep. I went to open an issue about it, but issues are not enabled on zeitkapsl/zeitkapsl…
score 1 · canonical
the reason I chose them was solely photos.
same, now they are trying to be everything. I discovered ensu via you and now I definitely don’t want to give them my money.
I went to open an issue about it, but issues are not enabled on zeitkapsl/zeitkapsl
indeed I would like that too, there usual way of submitting reports is through email on their support page.
score 1 · canonical
indeed I would like that too, there usual way of submitting reports is through email on their support page.
Yes I sent them an email with a few queries.
score 1 · canonical
-
why did they rewrite more than 6,000 lines in 3 days after my email?
Oh no, that sounds like slop was involved.
score 1 · canonical
I don’t really understand the problem here because fairscan is under GPL and all Ente code appears to be AGPL, which is allowed. Ente also did this quite openly (the blog post says that they even contacted the author).
So, as the title implies, it is only a matter of putting the attribution into the code itself. But as the code as been rewritten in a new language, I can understand that this has been removed without really thinking about it. They probably rewrite it in Rust whith Fairscan code next to them but, then, in all honesty, it is not the same copyrighted code. It is also well possible that the Ente engineer was using several academical sources and, as such, don’t remember that he had one main source for several functions.
What would makes OP happy? Having his name in the code itself? Or elsewhere? It is not really clear to me.
But, yeah, Ente could be a bit more thankful and handling this more elegantly. It might be that Ente people didn’t really understood what OP was really asking (I mean they talked about Fairscan on their blog, they contacted the author, …).
BTW, Fairscan looks awesome. I wonder how it compares to OSS Docs Scanner.
score 3 · canonical
Hi ploum,
Small technical note: GPLv2-only isn’t compatible with the AGPL; FairScan is GPLv3, so that isn’t the issue here.
The actual problem is attribution. It’s not a question of tone, it’s a question of license compliance: GPL code has to be declared, the copyright and license notices have to be kept, and any modifications have to be clearly marked and maybe dated.
What would makes OP happy?
Not the correct question. The question should be “How can they be compliant ?”.
score 1 · canonical
I don’t really understand the problem here because fairscan is under GPL and all Ente code appears to be AGPL, which is allowed. Ente also did this quite openly (the blog post says that they even contacted the author).
Opensource is not public domain. It frustrates me how often people act like it is.
The entire point of all of the open source licenses is that they are protected by copyright. Anyone can distribute a derivative work, but they can’t remove the original copyright owner, nor claim to be the copyright owner. Similarly they can’t relicense beyond what the original license permits, and even if they do, they still cannot remove the copyright information because the original author[s] have not rescinded their ownership of the code.
But, yeah, Ente could be a bit more thankful and handling this more elegantly.
What Ente did was steal open source code and claim ownership of code that was not theirs to own. They cannot claim ownership of derived code.
score 1 · canonical
IANAL I think beyond “having the name in the code itself”, GPL has specific, legal requirements. https://opensource.stackexchange.com/questions/4577/does-gplv3-require-attribution
The question is whether these AI rewrites skirt around this by being a “clean rewrite”. Again, IANAL, but the proponents seem to argue that it is, while the detractors say that it is not a clear rewrite, or it may not be a clean rewrite in specific circumstances.
I’ve personally written GPL code, and then had a company who is selling a service come and ask for features, all the while never making any changes they have made to the codebase freely available back. It sucks, because I spent probably upwards of 1000 hours of work on it, all for no financial compensation, while the company is making money off of my work without giving me a single cent. Which they’re allowed to do, but again it sucks.
I’d like to continue to write the software, but after my paid work I have hardly any energy to work on it, and also, what’s the use? Not only am I never going to see a cent from it, but also the people who use it are making bank off of my free efforts. Additionally, LLMs are trained on it and used to then argue that I should be fired from my paid job as what I do can be replaced with AI slop.
Bit of a rant!
score 1 · canonical
The question is whether these AI rewrites skirt around this by being a “clean rewrite”. Again, IANAL, but the proponents seem to argue that it is, while the detractors say that it is not a clear rewrite, or it may not be a clean rewrite in specific circumstances.
Because LLMs count as tools under current case law (as weak as it is), if you end up in court you’ll have to argue against the fact that:
- You saw copyrighted code
- Using a tool, you created substantially similar code
The current legal fiction of “having one agent describe what the existing code does to a separate agent that implements it” which is roughly like clean-room reverse engineering, just doesn’t work without the strict separation that requires multiple people to be involved - you are tainted, so you cannot have any influence on the output, which you inherently have if you’re the one controlling the LLMs.
score 1 · canonical
Look I’m not a lawyer. Your argument could be sound or not. In any case accepting these AI rewrites as clean room rewrites would basically render copyright worthless, and I doubt the courts will stand for that. Although I could be wrong
score 1 · canonical