Kimi overthinking is really expressive
K3 overthinks heavily but it can do double base64 decode with full error correction. it thinks this test is cute. I think they should ask why Sonnet 3.5 could do almost as good without CoT.
Is Kimi the first model whose reasoning tic is "Hmm" instead of "wait"?
Users dismiss Kimi AI's overthinking on binary-to-ASCII conversion and its odd 'Hmm' reasoning tic as dizzying, excessively convoluted, and weird.
Based on 3 visible X reactions from 5 accounts; directional sample.
Ask a question below.
Published answers will appear here.
K3 overthinks heavily but it can do double base64 decode with full error correction. it thinks this test is cute. I think they should ask why Sonnet 3.5 could do almost as good without CoT.
Is Kimi the first model whose reasoning tic is "Hmm" instead of "wait"?
Kimi overthinking is really expressive
Users dismiss Kimi AI's overthinking on binary-to-ASCII conversion and its odd 'Hmm' reasoning tic as dizzying, excessively convoluted, and weird.
Based on 3 visible X reactions from 5 accounts; directional sample.
Ask a question below.
Published answers will appear here.
