Appreciate the effort but it's unlikely even a single person will respect it given Bluesky is developing their own AI consent signal and barely anyone is using this. Feels like a kneejerk reaction to me ngl and not fully thought out

I want feedback from anyone who wants to use these to study the platform - should I include more data, like anonymized DIDs to allow reply network tracking, post interactions, etc?

I'm not going to argue on any of your legal assertions because I'm certainly no expert there, but I know for certain someone will collect this data either way.

Letting it happen in the open and working with us on privacy standards gives you more control over the process.

I think you're attacking the wrong people here. Would you rather us do this in the open, where the data is often anonymized and opt-out is available in many cases, or collect all data in private?

This is ultimately the kind of usage ATproto and the firehose were designed to support.

I'll be releasing bigger datasets in the future and intend on making it easier to opt-out when that happens. I want to make my datasets both useful in terms of research and privacy-respecting.

Great work, props for using a permissive license!