It does not, this is not in the law. This is a false misunderstanding that somehow keeps getting propagated. A few civil rights people have looked into it and don’t see it as a likely problem, and there haven’t been cases where the government was able to use this argument.

You would be aghast to spend any time in tech these days. I would wager more than 20% of the people in this field right now are wearing devices which record everything they say in every conversation and ship it directly to a transcription LLM and into some other tools. Go to SV and I bet this number jumps to nearer to 30-40%

People do NOT want smaller phones, that should be abundantly obvious.

>everyone is the same as me

How do you detect the watermarks and how do you remove them. Code please.

Right now we use a special algorithm to rephrase the submitted text without changing or losing meaning. Once Anthropic documents how to detect the watermark itself work to remove it more surgically can commence.

I remember when I was young and every billboard in SF was some stupid web 2.0 startup, or SaaS, or Crypto Shitcoins. Now we're plagued with AI billboards and the world is ending!

At least I could usually tell what a web 2.0 company actually did. With AI nearly being a commodity at this point, AI billboards are like seeing ads for different brands of table salt.

There were cryptic billboards in the last 2 tech booms too (dotcom, web 2.0) - the one that comes to mind is the Google ones that were a puzzle that led to a recruiting funnel (2004: https://www.npr.org/2004/09/14/3916173/google-entices-job-se...)

There are a lot of Monaco billboards up now which made me laugh.

There's a 2bit quant on HF already at ~1TB

Just let it manage itself, what could go wrong! :)

Congrats on quitting!

I love that you've said this to me more than once.

There it is