Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude Code is steganographically marking requests

thereallo.dev

Post reply on HN
Print view View original

271–280 of 817 posts

Re: Claude Code is steganographically marking requests

#271

Post by dkersten »

anon373839 wrote: post #261
civet_java wrote: post #238

There are some commentors in this thread downplaying the severity of a service provider being less than transparent about exactly what their shipped tooling does on customer's machines. That the provider's business needs necessitate the this behaviour doesn't justify their lack of honest disclosure. That honest disclosure would render the solution to their problem useless isn't my problem. If anything, that they thou…

Dishonesty seems to be a core value at Anthropic. I find myself wondering how anyone could have confidence in them after their repeated breaches of trust.

I honestly find it crazy how many people trust them for their business needs. For a business, you want consistency and no surprises. With them you get exactly the opposite.

Re: Claude Code is steganographically marking requests

#272

Post by breakingcups »

MallocVoidstar wrote: post #193
nixosbestos wrote: post #161

Earlier quoted context omitted.

I am also really confused and annoyingly stuck on this. I understand that the model name might appear in prompts for distillation (I guess? "You are RipOffModelv2, learn from these responses from Claude")? I guess the only explanation is that there's a side-telemetry channel that still sends some data to Anthropic, regardless of ANTHROPIC_BASE_URL overrides.

There are a lot of companies reselling Claude to Chinese users. You use their base URL but it's still going to Anthropic.

To clarify why Anthropic wants to catch these parties: they save all session logs and sell them to other LLM firms (for distillation) and have been known to use stolen credit-cards to pay for the Anthropic accounts.

I'm quite all right with the first, not with the second of course.

Re: Claude Code is steganographically marking requests

#274

Post by ForHackernews »

superfrank wrote: post #164
meowface wrote: post #92

Value judgment aside: I am a bit surprised at how sloppily they did this. I think they could've achieved the same effect while decreasing the odds of detection via reverse engineering. (This field is known as "underhanded code", coined by the Underhanded C contest: https://www.underhanded-c.org . It's a little-known "art"; little-known for probably self-explanatory reasons. There are much cleverer ways of achieving o…

It's also possible that there are more in-depth detection methods and that this was just a cheap and easy first step that hasn't been removed because it catches a lot of less sophisticated bad actors. It's unlikely that this will stop a big AI lab from distilling their model if they're really determined, but A) it may be enough to stop a bunch of fly-by-night token resellers looking to make a quick buck and B) you ne…

> fly-by-night token resellers looking to make a quick buck

aka market competitors reverse-engineering for interoperability

Re: Claude Code is steganographically marking requests

#276

Post by sanderjd »

loufe wrote: post #203
VortexLain wrote: post #56

Codex CLI is FOSS, unlike Claude Code, so Codex is less likely to do things like that, and it's one more reason to avoid Claude Code and Claude in general. Hopefully, many eyes will be looking into Codex for malicious things like that.

Genuine question though, why would I care about this if I'm paying for a subscription and adhering to TOS. I'm very skeptical about their privacy policy, business practices, and so on, but am curious what the negative about this is. Seems like it would work to my favour as a customer pushing back any date of the cutting of subsidies. That said, these fraudulent proxies are helping Chinese labs keep up, which might be…

Because they could use (or maybe are already using) similar techniques to do things you don't approve of, without your awareness.

Re: Claude Code is steganographically marking requests

#278

Post by bel8 »

Quinner wrote: post #188
bel8 wrote: post #155

Earlier quoted context omitted.

if by announce you mean shove it somewhere in a pdf with hundreds of pages, yes

https://www.anthropic.com/news/claude-fable-5-mythos-5 This is not hundreds of pages and it gets its own bold headline section.

Wrong. That page is after they walked back because of the outrage. The original was on the PDF as reported in this article:

> If Claude Fable stops helping you, you'll never know

https://jonready.com/blog/posts/claude-fable5-is-allowed-to-...

HN post with 1k+ comments: https://news.ycombinator.com/item?id=48467896

Re: Claude Code is steganographically marking requests

#279

Post by gleenn »

civet_java wrote: post #238

There are some commentors in this thread downplaying the severity of a service provider being less than transparent about exactly what their shipped tooling does on customer's machines. That the provider's business needs necessitate the this behaviour doesn't justify their lack of honest disclosure. That honest disclosure would render the solution to their problem useless isn't my problem. If anything, that they thou…

Whether or not you find Anthropic's behavior bad, theybhave been very loudly stating the foreign labs have been distilling their models for a while now. This seems like an obvious response to me that would be a mechanism to make that obvious.

Re: Claude Code is steganographically marking requests

#280

Post by Melatonic »

meowface wrote: post #92

Value judgment aside: I am a bit surprised at how sloppily they did this. I think they could've achieved the same effect while decreasing the odds of detection via reverse engineering. (This field is known as "underhanded code", coined by the Underhanded C contest: https://www.underhanded-c.org . It's a little-known "art"; little-known for probably self-explanatory reasons. There are much cleverer ways of achieving o…

Or they are doing both and this is the obvious part. Sort of like installing a bunch of real security cameras alongside a few fake ones
Post reply on HN