Fun, but annoying when articles like this reference online resources like github repos and hackernews threads without linking to them anywhere
https://github.com/instructkr/claw-code Maybe the hackernews thread? https://news.ycombinator.com/item?id=47584540
So is “Claw Code” supposed to mean Claude? I don’t see any explanation of what the repo actually is.
As far as I can tell, claw code is the python reproduction of Claude code that the leaker/discoverer wrote up to avoid getting sued or at least taken off github for hosting their leaked code. I haven’t looked at the repo much so it might be very incomplete, but maybe it works!
Anthropic pulled the npm package within hours and issued a statement: the exposure was “a release packaging issue caused by human error, not a security breach.”
I’m sure they chose the words “human error” to also imply the error is not to blame on their LLM, which remains an open question after reading the article (and likely will forever, although at the very least it would seem like the LLM did not detect this mistake).
Yeah, I had a good laugh at this. Half of the commits I review are coauthored by Claude, a fact that I’m sure Anthropic is thrilled to claim, but this colossal fuck up was obviously the work of a rogue intern or something.
The clean-room approach created a novel legal puzzle. Gergely Orosz (The Pragmatic Engineer) observed that Anthropic faces a dilemma: a Python rewrite constitutes a new creative work potentially outside DMCA reach. If Anthropic claims the AI-generated transformative rewrite infringes copyright, it could undermine their own defense in training-data copyright cases - the same argument that AI-generated outputs from copyrighted inputs constitute fair use.
The only moral justification for an AI rewrite. Love it.
My guess they are not even gonna challenge the “clean room” rewrite legally: the damage is done and it’s not really gonna be mitigated if they manage to take down the rewrite.
