3 ms·
Great premise but can't really agree with the execution. Felt like this makes too many implicit assumptions about LLM capabilities and traps without differentia
by not-chatgpt 6mo ago
Great premise but can't really agree with the execution. Felt like this makes too many implicit assumptions about LLM capabilities and traps without differentiating enough between a smart human vs AI.
- lukasec 6mo agoSmart humans, or humans with LLMs, solving them is not a problem. Main filter is agents vs deterministic API-key-farming scripts. Traditional CAPTCHAs also leak in the other direction (our agents crack them consistently).