Eweek content and product recommendations are editorially independent. We can earn money when you click on links to our partners. Learn more.
Artificial intelligence does not only change the world – it is also soaked squirrels.
Washington Post journalist Jeremy B. Merrill, has created a machine powered by AI called Squirrel Soaker 9000. Backyard DIY construction uses image recognition, Cloud Computing and a raspberry pi to do one thing: sprayed squirrels of a bird feeder. He notes that it does not harm squirrels.
It’s absurd. It is strangely effective. And this is a perfect example of how AI tools are accessible and how they appear quietly in places beyond the data center.
Configuration: open source + a pipe
The system is built around a fairly simple premise.
- Every 30 seconds, a camera takes a photo of the bird feeder.
- The image is analyzed by a fast.ai model operating on AWS Lambda.
- If a squirrel is detected with more than 70% confidence, the system activates a water sprayer, records the video and records the meeting.
The results: good, not 100%
The precision is not perfect – approximately 86.6% precise, according to its creator, with the cave of occasional mourning taken in the cross fire – but it turned out to be sufficiently capable of considerably reducing the tours of feeders by squirrels.
“Fastai (an in -depth learning library designed for researchers and practitioners) makes things really very easy – seriously, 13 lines of code – to cause an AI model to learn to distinguish two types of images, giving it examples of two types of images”, ” Merrill wrote in his blog article.
Bord Ai in its purest form
The Squirrel Soaker 9000 strikes on the basic themes that apply to the adoption of corporate AI. It is an excellent example of IA Edge: light models deployed on a small material, making real -time decisions based on image data. No heavy calculation, no GPU clusters, no sets of massive data. Just a python script, a camera and a clear goal.
For companies exploring AI at the edge, the principle is valid: local inference, rapid reaction time and autonomy without the need for constant human surveillance.
More lessons to learn
The project highlights a key point for AI beginners: you don’t need flawless data sets or complex pipelines to build something that works. Quick iteration and “fairly good” entries can go very far.
In addition, projects such as Squirrel Soaker 9000 can be built to solve a very specific problem with certain wood creatures, but they highlight a serious point: AI is no longer only in the laboratory.
