METRICS

Why do LLM outputs get worse even when metrics stay stable? [pdf]
4 points | 0 comments

Number of tokens shouldn't be the only metric
1 points | 0 comments

Show HN: Pathetic – Java pathfinding: 10k concurrent paths in ~7ms
8 points | 2 comments

Ask HN: Why would we care about "extended time horizons" and LLMs?
5 points | 5 comments

If society had a scorecard, what would be on it?
2 points | 9 comments

Show HN: NyaayWatch – Observability layer for the Indian judiciary
2 points | 0 comments

Show HN: Spinal – Prod aware code review and validation
1 points | 0 comments