We found it pretty easy to build a little k8s controller for our own purposes to do this -- see https://news.ycombinator.com/item?id=48478994 . You probably don't need to implement this as a plugin or hook, pgdog supports dynamic reload of its configuration without dropping existing connections.
Although I'm the type to shy away from adding extra layers in my architectures when I can help it, pgdog has been an absolute breeze to use :)
Considering "I Can Has Cheezburger" and that kind of lolspeak is nearly 20 years old now, I'm not surprised to see some people unfamiliar with it nowadays :-)
We successfully did this with pgdog at $JOB using our own "controller" -- the same service that handles deploying new instances of our application (instancing an argoCD Application that fires Crossplane DB creation, making new Deployments of bricks, etc) will also, at the end of that process, scan the cluster for Database CRDs, use those to generate a new pgdog.toml + users.toml, update the Secrets in the cluster, enable maintenance mode on all pgdog pods, do a live config reload on each of them, then disable maintenance mode (this is to make the change atomic between all the pgdog instances). Downtime there is about 2-3 seconds and all it does is make new SQL requests from existing clients wait, it doesn't break the connection or anything.
The obvious AI headings, pointless genned image of people (I'm starting to think islam had a point with discouraging depictions of human figures), and especially the blurry, artifacted, distractingly skeuomorphic diagram, with random wire traces going everywhere... this is a technical blog, not an investor sales pitch! Every time I see one of these, I have to double-check for a second if I'm not on some phishing SEO site!
If even Google, previously a gold standard of technical writing, is falling prey to this kind of laziness, then I have nothing to worry about -- knowing how to write without a language model in the driver's seat is gonna be a top tier skill in the future...
A damn shame too, as I've been following the progress of JXL in the standardization pipeline for a few years now and was quite interested in the historical breakdown, but all that's gonna stick with me from this is the disrespect I felt as a reader.
... with a high-resolution scan of the work itself available for download, to boot. I really appreciate whenever museums go out of their way to share those publically! Much better than many paintings only officially available as some 400px thumbnail.
Oh hey, Tixl on HN! It's probably my favorite piece of FOSS creative software, and honestly blows After Effects out of the water for me (as a hobbyist who likes cooking up some vis for my DJ sets).
I'm not kidding when I say it has the potential to become the blender for 2D/3D VFX -- the node engine is super powerful, the primitive building blocks are all well thought-out and integrate with each other very nicely, the performance characteristics are amazing (all optimized for realtime!), and there's a ton of I/O for everything from mouse input to OSC/MIDI, camera control, elaborate audio reaction... and also just plain TCP/UDP/HTTP/Websockets! It's such a powerful glue piece, but also tons of fun to mess around with on its own.
The best part? You can create your own components, define your inputs/outputs, and compose them together. The even bestest part? You can dig into the predefined components/effects and see how they work, as they're very often implemented in the same way! The visual editor all drills down to C# in the end, and you can drop into the code or write some HLSL shaders if you want, all with hot reloading.
How relevant is this (and the NSA's general spying capability) in 2025?
We hear a lot about local agencies perusing the services of private companies to collect citizens' data in the US, whether that's traffic information, IoT recordings, buying information from FAANG, etc. What's the NSA's position in the current administration? (e.g. we've heard a lot of noise in the past about the FBI and CIA getting the cold shoulder internally. I wonder how this applies to the NSA.)
> Palantir is firmly cemented into military-industrial infrastructure, and business is booming, but Karp is not letting up. He has said he wants Palantir to be as dominant and indispensable as IBM was in the 1960s, when it was the world’s largest computing company and shaped the way government and private companies did business.
I was thinking even before this line that he gives off the impression of really admiring IBM, especially their German stint from 1933 to 1945.
I finally took the plunge and upgraded from Ubuntu 22.04 to 24.04 on my work laptop. With that come all the gnome package updates and such.
I reboot, load up my session, and a little while after need to grab a couple files from a zip archive. I double-click it in Nautilus, nothing happens. I give it a couple more clicks before I suspect something broke, right click it, and see "Extract" as... the default cursor action...?
I go back up and see five fresh copies of the folder that was inside the zip. I delete them all, go back to the file itself, right click, open with > file-roller. I try and drag'n'drop the couple files I need: doesn't work, for some reason. Great, they've broken drag and drop, too.
> is drag and drop extraction from Nautilus (or other file managers) really that important?
Why, yes it is! It's even worse if you go and _break_ drag'n'drop support on X11! I'm not even using Wayland!
But, oh well, looks like file-roller is unmaintained and outside of core Gnome scope now. Nautilus' zip capabilities are enough, they say! Why would a user want to inspect the contents of a zip archive before wanting to extract it, or god forbid select specific files, after all? Definitely not worth keeping as a core OS/DE feature.
And the PR on file-roller that fixes this on wayland with... a custom fuse virtual filesystem?! has been untouched for the past 2 years, never to be merged.
Otaku/doujin culture, and the creative industry that rose around it in Japan, is as good as it gets when it comes to finding good ideas and propping them up. Basically anyone who can draw can release their own manga/webnovel/illustrations on pixiv, twitter, and others, get a couple volumes out with this or that publishing house, see where it goes and whether the public catches on. Self-publishing plays a huge part in this, whether it's doujinshi (self-made [often derivative] books sold at conventions like Comiket, Comitia, or online, with a substantial proportion of r-18 [but not only] content), doujin music albums, indie games and visual novels, etc.
Funding for anime adaptations is plentiful, and fan support helps bridge the confidence gap where production committees (consortiums of multiple publishing/IP companies pooling the money to distribute the risk) won't go for more indie / experimental works. Profit is recouped on developing the brand and merch, while leaving plenty of room for directors and studios to establish their own auteur identity. Studios are getting leaner and more focused these days, splitting off into smaller entities kicked off with a project or two with more margin for talent to shine.
It's not all perfect, though. Freelance work and lack of mentoring has really put a dent into the supply of new animators, who lack job security and often swing between studios left and right, but there are industry efforts to fix this and preserve knowledge, with the oldest pioneer animators now starting to hit their 60's and 70's.
Question to anyone with experience in this domain: I have CSAM spam problems on a forum I host, with bots putting link shortener URLs embedded in images rather than the post body. Traditional OCR software deals poorly with them due to font modifications and intentional text edge modifications, and I'm obviously not gonna use a SaaS/closed source model to upload a bunch of may-be-may-not-be-CSAM pictures, so looking for a way to do this locally, with cheapish inference if possible (I don't mind spending a minute of compute to get the result out for one image, but need to do it on the CPU).
Is there any small model that would do this effectively, with pure text extraction (without going for any kind of formatting or whatnot)?
Been using it for the past few years, nothing bad to say about it, lovely piece of software. Vendor lock-in is very present in this field, with different brands of controllers supported by a myriad of proprietary DJ applications all more interested in onboarding you to their music subscription services rather than implement useful features or support open protocols.
Meanwhile, Mixxx allows you to write your own adapter scripts for any controller you have (as long as it outputs MIDI), and there's a built-in library featuring scripts for the most common commercial controllers and MIDI devices out there.
Might be wise to make it more apparent that you need to close the right sidebar before you can do anything on-screen -- I tried clicking on the "Get Started" buttons, found them to be unresponsive / non-interactive, tried clicking on stuff on the side, and closed the page at first, thinking it wasn't tested on Firefox.