If they scrape the updated comments again and ingest copyrighted text, you are poisoning the data.

I think you missed the part where you were strongly suggested “not” to use copyrighted text.

The point is not to get rid of the original text. It’s to “poison” the training data.

Nice write-up!

It’s been crazy to see the AI hype train amusement park ride lately.

It’s not all bad! The Cleon part is pretty cool. You know… the part that is not in the books at all.

Thanks!

Could you expand a bit on how your setup works? I think I get the big picture, but not sure how tailscale fits into it.