← Back to radar

Observation record / Hacker News

Using LLM for confidence scoring is a bad idea

First seen 10/22, 10:45 · Last observed 10/22, 12:05 · Source data is current

Read the original source ↗

Title-based outlineUsing LLM for confidence scoring is a bad idea

Why it mattersReal snapshots show a rank rise within this source, indicating changing discussion, not verified facts.

Published 10/21, 12:27

30 minutesRank movement0

Baseline observed 10/22, 11:35

1 hourRank movement+5

Baseline observed 10/22, 11:05

24 hoursRank movement—

Building history

Source rank history

9 observations

Momentum 36/100 = max(0, 1h rank gain) ÷ baseline rank × 100. Platform metrics are not combined.

91710/22, 10:45 · 17 places10/22, 10:55 · 12 places10/22, 11:05 · 14 places10/22, 11:15 · 10 places10/22, 11:25 · 10 places10/22, 11:35 · 9 places10/22, 11:45 · 9 places10/22, 11:55 · 9 places10/22, 12:05 · 9 places10/22, 10:4510/22, 12:05
Official public API · linked titles only· Author jimauthors
View original observation times and ranks
Observation time (Beijing)Source position
10/22, 12:059
10/22, 11:559
10/22, 11:459
10/22, 11:359
10/22, 11:2510
10/22, 11:1510
10/22, 11:0514
10/22, 10:5512
10/22, 10:4517
Daily aggregate history
日期Best daily positionLast daily positionObservations
2026-10-22999