#10942.0AIT2 — media / chartCooling

After 600 experiments, here's what I discovered about JEPA world models

Spotted 2026-11-14 03:07 (UTC+8) · original event 2026-11-11 · sources: Hacker News · spreading on 1 platform

Heat score

Rank score: 6040% — where it sits on the charts
Discussion score: 8040% — how much people are talking about it
Cross-platform score: 3020% — how many platforms it spread to
Composite heat score:42.0rank × 0.4 + discussion × 0.4 + cross-platform × 0.2 — includes a -20-point cooling decay (peak passed, attention fading).

Why it matters

Everyone argues about what a world model should be, but few people have actually run the experiments - 600 of them is a rare, concrete data point in a debate usually fought with opinions and press releases.

The angle

One person ran 600 experiments on JEPA world models - here's what actually works, and what the generative crowd gets wrongSuggested format: Short video · for AI researchers, ML engineers, world-model followers

Evidence chain

• HN #2, 55 pts/5 cmt (2026-11-14)• Haddy Alchaer personal blog 2026-11-11: 600 experiments on JEPA world models, on the debate over what a world model should be - generative models that predict the future by generating it in pixels vs JEPA's predict-the-representation approach• 2026-11-14 04:07 update: HN #2, 119 pts/25 cmt• 2026-11-14 05:07 update: HN #4, 154 pts/30 cmt• 2026-11-14 06:07 update: HN #4, 174 pts/31 cmt• 2026-11-14 07:08 update: HN #8, 191 pts/32 cmt• 2026-11-14 08:08 update: HN #6, 200 pts/33 cmt• 2026-11-14 09:07 update: HN #9, 209 pts/35 cmt• 2026-11-14 10:07 update: HN #12, 217 pts/36 cmt• 2026-11-14 11:07 update: HN #14, 219 pts/36 cmt• 2026-11-14 12:07 update: HN #13, 225 pts/37 cmt• 2026-11-14 13:07 update: HN #15, 231 pts/37 cmt• 2026-11-14 14:08 update: HN #12, 238 pts/37 cmt• 2026-11-14 15:07 update: HN #12, 244 pts/40 cmt• 2026-11-14 16:07 update: HN #14, 250 pts/41 cmt• 2026-11-14 17:07 update: HN #14, 257 pts/41 cmt• 2026-11-14 18:07 update: HN #14, 262 pts/41 cmt• 2026-11-14 19:07 update: HN #15, 264 pts/43 cmt• 2026-11-14 20:07 update: HN #17, 267 pts/45 cmt• 2026-11-14 21:07 update: HN #17, 271 pts/45 cmt• 2026-11-14 22:08 update: HN #20, 273 pts/45 cmt• 2026-11-14 23:07 update: dropped out of HN top 30• formula 68.0, stored 48.0• 2026-11-15 00:07 update: returned to HN top 30 at #30, 278 pts/46 cmt• formula recomputed 62.0 (old formula 68.0), cooling cleared• 2026-11-15 01:07 update: dropped out of HN top 30• formula 62.0, stored 42.0