Live data from Hacker News

Claude Code is steganographically marking requests

thereallo.dev

1–10 of 817 posts

Re: Claude Code is steganographically marking requests

#3
This is very interesting. Combating resellers and distillation seems like a very difficult problem indeed. Interesting to me is that these techniques mentioned in the article are just like anti-observation techniques used by some of the more sophisticated malware out there, however defeating them is pretty trivial.

Re: Claude Code is steganographically marking requests

#7

“So the feature mostly punishes the exact people who are easier to fingerprint: normal developers doing weird but legitimate things” What’s the punishment here exactly?

Returning invalid poisoned different results that were not what you paid for

Re: Claude Code is steganographically marking requests

#8

“So the feature mostly punishes the exact people who are easier to fingerprint: normal developers doing weird but legitimate things” What’s the punishment here exactly?

Output poisoning and/or eventual account bans, if I had to guess.

Re: Claude Code is steganographically marking requests

#9
post #3

This is very interesting. Combating resellers and distillation seems like a very difficult problem indeed. Interesting to me is that these techniques mentioned in the article are just like anti-observation techniques used by some of the more sophisticated malware out there, however defeating them is pretty trivial.

seems ironically like a similar problem of content owners trying to filter bot scrapers from legit users

Re: Claude Code is steganographically marking requests

#10

“So the feature mostly punishes the exact people who are easier to fingerprint: normal developers doing weird but legitimate things” What’s the punishment here exactly?

They probably run a heavily dumbed down version of the model, same as what they got caught doing with Fable.

And that's also why, as a legitimate customer, want none of it, you never know if you accidentally entered a zone they don't like.

Post reply on HN