ghxst

The proxy you suggested sounds similar to a WAF, I don't doubt there's use for it but I would assume it comes with similar downsides.

My fear is that platforms that will go to great lengths to enforce this will become an RL playground for some devs to train their chatbots.

I would say you can get to this point far below 100 million people, especially on web. Some people are truly special and have some kind of setup you just can't easily reproduce. But I agree, you do really have to be confident in your ability to control rollout / blast radius, monitor and revert if needed.

How does it compare against tshark?

Any docs on this? I didn't know this was a thing.

The use of "storage.googleapis.com" is probably because it's an "authority" domain that apps can't easily ban without side effects. Buckets can typically be used as a static site host where u can host a client side redirect, depending on how you set it up you can make it almost impossible for an app to ban a campaign in real time.

I'm working on a shared remote box for AI assisted development, will definitely look at this for some inspiration.

Don't forget color bots!

Most of them abuse the ip pool attached to lambda from my experience.

They do KYC when you want to unblock certain domains.

That really sucks. Any advice on how to "negotiate properly" to avoid a situation like this?

Is this constitution derived from comparing the difference between behavior before and after training, or is it the source document used during training? Have they ever shared what answers look like before and after?

If you go http native, could you leverage range headers for offsets?

I've used wireguard for a while, not sure why I never considered doing BGP over it, might make for a fun weekend project.