|
Phase 1 post-mortem
|
|
1
|
108
|
August 11, 2026
|
|
Phase 1 write-up: Stabilizing cumulant propagation at depth 32 — a trajectory-calibrated moment chain, with an error budget (submission #314695)
|
|
3
|
1196
|
August 11, 2026
|
|
[Phase 1 write-up] Where the closure family's floor actually is, and what the residual term costs per call (submission 324409)
|
|
0
|
107
|
August 11, 2026
|
|
[Phase 1 write-up] Metered dead-compacted Kerdock cubature with exact compiler reductions (submission 326680)
|
|
0
|
123
|
August 11, 2026
|
|
Phase 1 Write-Up: A Modern Approach
|
|
0
|
104
|
August 10, 2026
|
|
[Phase 1 write-up] The non-Gaussianity wall at depth 32, measured with certificates — and a falsification-ledger methodology (submission #326094)
|
|
0
|
136
|
August 10, 2026
|
|
Phase 1 submission selection & Phase 2 Launch date
|
|
3
|
299
|
August 10, 2026
|
|
Smoke wall time limit is different from evaluation limit?
|
|
1
|
334
|
August 10, 2026
|
|
[Phase 1 write-up] Fixed grid of directions instead of random sampling (submission #320802) and five things that did not work
|
|
0
|
140
|
August 10, 2026
|
|
Has anyone else noticed that the 50 leaderboard MLPs are unusually easy / low-variance?
|
|
4
|
273
|
August 10, 2026
|
|
Phase 1 update: flopscope v0.10.0, cost-model fixes, residual-time safeguards, and updated deadlines
|
|
9
|
1118
|
August 10, 2026
|
|
[whest run] Residual Wall Time in the budget-breakdown panel reads ~39,807 s under --runner subprocess
|
|
1
|
441
|
August 10, 2026
|
|
Cost-model question: releasing temporaries of counted ops is charged to residual — and a request to expose out= on fnp.matmul
|
|
1
|
496
|
August 10, 2026
|
|
Phase 1 nomination interface and the Aug 17 write-up — where do they live?
|
|
1
|
368
|
August 10, 2026
|
|
Rules clarification needed on compute budget exceedance
|
|
3
|
384
|
August 10, 2026
|
|
Request to extend the deadline of Phase 1
|
|
10
|
1130
|
August 9, 2026
|
|
[flopscope 0.10.0] ALL of `flopscope.stats` silently promotes float32 to float64, doubling the bill downstream
|
|
1
|
449
|
August 7, 2026
|
|
[flopscope] budget_summary_dict() cost grows with process history — silently inflates residual and can cost the 0.1 multiplier floor
|
|
1
|
587
|
August 7, 2026
|
|
Recommendation: restoring the estimation framing (neutralizing the wall-time compute channel)
|
|
1
|
615
|
August 7, 2026
|
|
Potential flopscope accounting bypass bug
|
|
3
|
935
|
August 7, 2026
|
|
How exactly is phase 1 submission validation going to work?
|
|
1
|
300
|
August 7, 2026
|
|
A bet on a floor, filed before the reveal
|
|
4
|
857
|
August 7, 2026
|
|
Will the private re-evaluation verify instrumented share?
|
|
1
|
403
|
August 5, 2026
|
|
Grading queue backlog since the v0.10.0 rollout
|
|
2
|
288
|
August 4, 2026
|
|
Rules clarification: are operations on the NumPy-backed arrays reachable through flopscope.numpy inside the intended accounting boundary?
|
|
5
|
665
|
August 3, 2026
|
|
Phase 1 final evaluation question: Overfitting to public instances
|
|
3
|
710
|
August 3, 2026
|
|
Urgent: submission limit is rolling 24h, not the published UTC-day reset
|
|
1
|
325
|
July 31, 2026
|
|
Phase I Submission: Structure-Aware Estimation for Random ReLU MLPs
|
|
0
|
887
|
July 27, 2026
|
|
There is a bug in flopscope.numpy.linalg.solve that undercounts batched right-hand sides
|
|
5
|
850
|
July 25, 2026
|
|
Submissions score 50/50 with 0 failures but flagged "Evaluation could not complete"
|
|
1
|
449
|
July 23, 2026
|