| *** n2klb_ is now known as n2klb | 02:42 | |
| abranson | Mister_Magister: nice work. that gst agents.md is very well written. I'd agree with all of that, except maybe the commit messages. They should be reviewed and kept concise, but an LLM is often really accurate with its information with those and that's really important for future blame and bisect. | 13:59 |
|---|---|---|
| abranson | Mister_Magister: would be great indeed if they do fix that. I guess PCs would also see this if the machine suspends when a video player is open? | 14:00 |
| abranson | Mister_Magister: you're right about the qtwayland PR. will have to ping somemal about that. everyone's v busy right now. | 14:01 |
| titinonero_ | Hello all, is OBS down? | 14:04 |
| *** titinonero_ is now known as titinonero | 14:59 | |
| Mister_Magister | abranson: yeah i really like gst's LLM policy it sits well with me | 16:03 |
| Mister_Magister | abranson: also thanks for replying i thought you've forsaken me :D | 16:03 |
| sharks[m] | Do we have any status updates yet on the OBS ddos? I suspect its some AI scraper gone mad, but you'd think it would have calmed down by now. OBS has been unreliable at best for nearly two weeks | 21:02 |
| mal | yeah, there has been a lot of scrapers of something causing issues for it, not sure what changed that caused it to happen recently | 21:12 |
| x2s | probably something like this: https://people.kernel.org/monsieuricon/creepy-crawlies | 21:16 |
| x2s | sucking up everything, giving back nothing | 21:17 |
| mal | yeah | 21:20 |
| mal | you would assume it would be more efficient to get local clones of repo and just sync those when needed | 21:23 |
| mal | and faster | 21:23 |
| mal | and scraping via cgit seems just stupid | 21:27 |
| attah | something something cloudflare? | 21:28 |
| sharks[m] | I don't really understand what of value to an LLM is on OBS. It's just a collection binaries with links to git repositories, no? | 21:28 |
| sharks[m] | I agree attah, perhaps cloudflare or anubis or something is needed | 21:28 |
| mal | I think you gave get the package sources from there also, but that is just stupid way to do that, except when the package is so called dumb package which doesn't use git | 21:29 |
| mal | *can get package sources | 21:33 |
| sharks[m] | x2s: That's an interesting article, thanks | 21:40 |
| Renaud[m] | I had to restrict public access on my self-hosted forge because a bot from meta was using the web ui to do git blame on everything it found, and I mostly had clones of some big repositories already available on github… | 21:40 |
| sharks[m] | I don't think restricting public access to OBS is a good idea, as it's a key piece of the puzzle for users verifying what is actually installed on their phones. | 21:47 |
| sharks[m] | It would be nice if osc still worked from the command line, so even if the AI scrapers took down the web interface we could still get some work done | 21:48 |
| Renaud[m] | I wasn't saying we should restrict public access to obs, I was just pointing out the fact that ai scrapers don't really care for doing the right thing | 21:54 |
| Renaud[m] | It just sucks :/ | 21:55 |
| x2s | And think about all the great applications that are now generated and will be unmaintained in two month... | 21:57 |
| sharks[m] | It does suck. And there's no easy solution, but surely something like Anubis can't hurt. I'll gladly wait a few seconds for a page to load if it means it does have a better chance of actually loading. | 22:01 |
| piggz[m] | +1 ... i think Keto already tried anubis, but maybe worth another go, i guess its just time as usual | 22:11 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!