personally I find those experiences really funny especially in my life. looking back I think most people find humor in it, i could be wrong? I don't think so though
Fair feedback. Definitely more backlash than I expected. The intent was to experiment with large-scale analysis, not add noise or put strain on shared resources. I’ll be more thoughtful about this kind of thing going forward.
Totally fair callout. I should’ve been more careful here and leaned on the provided datasets / bulk access instead of pulling things at scale. That’s on me.
I’ll make a donation to support the project regardless. Appreciate you raising it.
The pet detection part isn’t the point, that’s just a visible output. The actual goal was to stress test agents + distributed compute on something non-trivial.
One of the Burla founders here. Not a joint project with Airbnb. I’ve been experimenting with giving agents access to Burla clusters and letting them run with analysis ideas I find interesting. This was one of the results.
The branding is a bit much, fair call, but the intent here was just to explore what these agents can actually build when you give them access to large amounts of compute.
I'm going to publish an Airbnb example tomorrow where I scraped 1,406,718 photo URLs from public listing pages. For that I used https://docs.burla.dev/ which is a high-performance parallel processing python library I've been working on for a few years now.
I fetched every Met Museum CC0 artwork with a published image, generated CLIP embeddings for all 191,922 of them, and searched for visually near-identical pairs separated by 2,000+ years. I used Cursor (Opus 4.7, max effort) to help build it, including spinning up a ~500 vCPU Burla cluster for the image downloads and embedding jobs. I also had it study popular museum websites for design inspiration while building the site. The full pipeline took about 50 minutes.