+0.38 Human coders are still better than LLMs (antirez.com)
655 points by longwave 277 days ago | 735 comments on HN | Moderate positive ~lite vlite-1.4
Summary ~lite Technology AI Acknowledges
Author discusses human coder superiority over LLMs
EQ 0.60
SO 0.70
TD 0.50
Lite evaluation by llama-4-scout-wai · editorial channel only · no per-section breakdown available
Longitudinal · 3 evals
+1 0 −1 HN
Audit Trail 8 entries
2026-03-02 12:49 model_divergence Cross-model spread 0.38 exceeds threshold (2 models) - -
2026-03-02 12:49 eval_success Lite evaluated: Moderate positive (0.38) - -
2026-03-02 12:49 eval Evaluated by llama-4-scout-wai: +0.38 (Moderate positive) 0.00
reasoning
Editorial discussing AI vs human coding capabilities
2026-03-02 12:44 model_divergence Cross-model spread 0.38 exceeds threshold (2 models) - -
2026-03-02 12:44 eval_success Lite evaluated: Moderate positive (0.38) - -
2026-03-02 12:44 eval Evaluated by llama-4-scout-wai: +0.38 (Moderate positive)
reasoning
Editorial discussing AI vs human coding capabilities
2026-03-02 12:41 eval_success Lite evaluated: Neutral (0.00) - -
2026-03-02 12:41 eval Evaluated by llama-3.3-70b-wai: 0.00 (Neutral)
reasoning
Technical post with neutral rights stance