With artificial intelligence being the bête noir of the moment, there are some projects using it which maybe don’t bring much to the table. So it’s nice to see one that uses it in a creative way, and delivers something new. [arnegiacomo]’s e-paper screen is a great example, as it draws a picture in real time of whatever birds are outside.
Behind the quite large screen sits a Raspberry Pi 5, and on that is BirdNet-Go, an AI-powered birdsong classifier. A USB microphone catches the birdsong, and Birdnet comes up with the species. The birds on the display are then those species as pictures from 19th century bird spotters guides, assembled into a collage. You can even see what the current set of birds it hears are, live, and they are a representative cross section of the European birds you’d find in Norway where it’s located.
We like this project, both for the bird book vibe it gives, and the creative use of machine classification. Surprisingly this isn’t the first project in this field we have seen over the years.
Oh, that’s a really nice project!
This is very classy. I could never care about birds but I appreciate the project. I have a friend who built a Raspi device running birdnet and installed it in a nearby forest
Correction, I do care about birds, but only as long as they’re in my plate!
Also similar to https://hackaday.com/2026/08/08/track-bird-visitors-with-a-raspberry-pi-and-a-usb-mic/
Was thinking exactly the same!
“hhuuuh, 3 weeks ago we had this covered already”
You’re confusing one type of machine learning system for a different, worse one. Both of these projects are harmless.
This <3
Maybe. BirdNet is probably “harmless”.
I think that [Arne] probably used LLM code-writing tools (looks like Claude?) to produce / modify / test the code that runs this project.
OTOH, [Arne] appears to be a coder by profession, so it’s likely that he reviewed it and made sure that it makes sense? But who knows.
Hi, [Arne] here. Can confirm my dayjob is coding as a software dev and that I’ve used AI in assisting development in creation of this project. Both Anthropic’s Claude and some different models through OpenCode Go have been used to write code, testing and documentation. Can also confirm that I’ve reviewed all code, but cannot promise that it always “makes sense” – in the sense that code doesn’t always just “make sense”. Code-wise this project is quite simple – most of the work has gone into deciding on hardware, features and most of all sourcing art.
Happy to answer any other questions or critiques :-)
Cuts to Hitchcockian / Black Mirror-esque scene, with AI communicating with birds to spy on and attack humans.
Hi, original creator here. You can of course roast both my code and writing as much as you’d like (feel more than free to) :-) However I’d like to just specify that the project itself runs fully local on a rpi, without any gen-ai. BirdNET is a classification model created by Cornell Lab of Ornithology and runs locally on the pi via BirdNET-Go. The images/illustrations are cut from scans of hand drawn natural history plates – basically old book scans from the 1800s found mostly on wikimedia commons (all artists and sources attributed in the source code).
AI in the form of LLMs has been used as a tool to aid development and writing the source-code, but the core of the project does not rely on it. Happy to answer any other questions you may have about it or my views regarding AI :-)
RE: very nice project and thank you for sharing with us.
On a lighter side of things (not trolling) may I suggest adding “How to Tell Birds from Flowers” by certain Robert Williams Wood, nice 100 year old parody that would hopefully add humorous dimension to the project. Can be safely ignored : ]
Definitely similar, but I appreciate that this one showcases human-drawn art rather than having a diffusion model create each bird picture. The technical writeup for the other post is very nicely done though; I like how it explains the packing algorithm.
You could combine it with a nerf minigun to fight birds which are destroying insects.
It’s “bête noire” (literally black beast) not “bête noir” (which means nothing or could be understood as “stupid black person” or “common black color”)
Why would you do that? They all belong in the ecosystem and it is not our place to interfere any more than we have.
?
I have an acquaintance who has a similar setup.
Apparently it identifies other wildlife too, but labels them as birds…
https://github.com/user-attachments/assets/7d6e9d49-38bf-4df7-8f63-e52b13b579c1
I’d like to point out the distinction between automatically drawing a picture and “automatically”(if you want to call spectral analysis automatic) cueing a pre-drawn picture that has been associated with that waveform. But anyhow, bravo.
(“Bête” is feminine, so “Noir” (the (absence of) color) takes a final e : “bête noire”)
Now its THESE kinds of things that are going to make AI into a truly new medium. All media mediums start out emulating the mediums that came before it. Vaudeville turned into movies and then tv that had vaudeville type shows and that evolved into sitcoms and dramas, same with art. I reas an article that said all of this much more elegantly lol, i can’t remember the specifics but thats the thrust of it
Now its THESE kinds of things that are going to make AI into a truly new medium. All media mediums start out emulating the mediums that came before it. Vaudeville turned into movies and then tv that had vaudeville type shows and that evolved into sitcoms and dramas, same with art. I reas an article that said all of this much more elegantly lol, i can’t remember the specifics but thats the thrust of it
It’s all fun until you see Mothman
Large numbers of mostly silent Canadian geese leaving their poo on the school grounds everyday. Can’t harass or annoy them either because of migratory bird act. The other day with full practice going on all the brass band instruments were sitting on the side lines where they poop. Finger sized! Not gonna toot that horn! We need robot “dogs” to chase them off before they land.