Then this might be the reason:
I absolutely adore the behavior of AI companies. DDOSing infra for scraping.
Illegally downloading copyrighted content for training.
Openly requesting that copyrighted content should be free for them “or their business model wouldn’t work”.
Oh and now the public should also build them AI datacenters - best free of charge for them.
I wouldn’t be mad if we put that app in front of the forum.
@darix
I won’t ‘heart’ your post. Their arrogance just makes me angry and sad.
If anyone is interested, the anime girl in question is GitHub - TecharoHQ/anubis: Weighs the soul of incoming HTTP requests using proof-of-work to stop AI crawlers
gnome’s gitlab instance is already using it, and I’ve been seeing it pop up more and more.
It seems to also block search engines from indexing you. It seems like a horrible solution and counter productive to the goals of this forum.
It’s a good solution for a lot of services but I wonder the effect it would have here, especially when it comes to getting new users.
Interesting because I’ve just been told the following on another site:
I agree, but it may come to this if no other defense is found. Forums like this one are probably not yet targeted to a large extent because harvesting content from other places is a higher priority, but it will arrive here too eventually.
That said, the most toxic example in the article linked above is submitting LLM-generated nonsense bug reports to FOSS projects with bounties, expecting that one of them will get through. This just ties up crucial developer time, and makes it difficult to get actual bug reports through. If a malicious actor wanted to wreck key software infrastructure, this is one thing they would do.
Yeah of course, if there’s no other solution and it starts bombarding pixls.us with huge bandwidth costs, it is what it is. I guess user outreach can be improved by other means besides search engines.
Lots of FOSS communities already suffer from uniform userbases that end up forming echo chambers, would be a shame for this forum to become like that.
There is no danger of that as long as we can invent new tonemappers 
Well you can also start talking to your representatives in your parliaments and demand that AI companies should be regulated. because their current free for all attitude in all regards is the problem. yes tools like anubis will probably just delay them I also wouldnt wonder if they start implementing solving the JS puzzle and then we are back to square one.
so if you want to help with the real solution: speak to your representatives and make your voice heard. depending on the countries you live in, this might also be an important step to take for other reasons.
Can we conclude that “other traffic” is also crawlers? If so that’s an insane amount relative to real users.
I would guess yes, because AI crawlers are known to use residential IPs and not send a user agent.
The reason we don’t feel it so much is that we have a bit of overhead on our server.
Ignorance
Malicious compliance, we let them in but we serve the code sorted.
For my own personal infrastructure, which is also getting railed by LLMs, I’ve been looking at https://iocaine.madhouse-project.org/
I tried Anubis and it only shows up for less than half a second, and only the first time you visit. In that time it calculated over 57,000 hashes satisfying the proof-of-work requirements.
What’s the unit on the y axis?
Page views. X axis is the date
IRL LOL
