alt Hacker NewsI hear these critiques a lot, and whilst I'm inclined to agree, I have never had an answer to: how do you test that ? Basically a captcha most humans can pass, that an LLM never will ?
I hear these critiques a lot, and whilst I'm inclined to agree, I have never had an answer to: how do you test that ? Basically a captcha most humans can pass, that an LLM never will ?
We could call it the
The Universal Reasoning In Novel Groundings
test