I expect a drawn-out dilemma with a backroom deal. Else, a stalled award. Exciting times!
On Wed, 09 Sept 2026, 10:13 SunTzuWarmaster, <[email protected]> wrote: > I imagine all the mathematicians in the room for the meeting: > > "So we have the choice of giving it to the AI company or the mathematician > whose data they stole." > ... > "We math people agree the math prize goes to the math person, right?" > > On Wed, Sep 9, 2026, 09:06 Quan Tesla <[email protected]> wrote: > >> Technically, OpenAI exploited AI speed to be the first to submit a 166 >> page complete thereom using a different approach. In a narrow sense, that >> gives them the prize advantage. >> >> However, in a more general scientific and IP sense, if AI didn't ingest a >> "benchmark" model, it would probable not have attempted an "alternative" or >> "different" approach. Therefore, the actual date of "prior art" proof moves >> further back. >> >> A probably caused B, the former had to exist for the latter to emerge. >> Therefore, the former existed a-priori. >> >> This is not a patent application. The branches may have physically borne >> fruit first, but the evidence of an ingested root-recipe stemmed from the >> nutrient uptake. >> >> Chickens and Eggs >> >> On Wed, 09 Sept 2026, 09:40 SunTzuWarmaster, <[email protected]> >> wrote: >> >>> There's potentially a perfect ending to this story - it's that the two >>> human authors get the prize money and openAI gete a funding round. >>> >>> I believe that humans currently award the prize money, right? >>> >>> On Wed, Sep 9, 2026, 06:28 Quan Tesla <[email protected]> wrote: >>> >>>> Thank you for sharing. This is just "wow!", but no surprises, really. >>>> >>>> It's not true that collaborating with a public, or subscribed-to LLM >>>> auto-invalidates IP claims. Privately, LLMs vlaim confidentiality and >>>> limited historical memory span. As my recent research report proves, this >>>> seems to be a convenient false fabrication. >>>> >>>> E g., recently, a public, non-subscribed to AI quoted me in a >>>> conversation we had back in March, 2025. Selective memories. >>>> >>>> However, full conversational transcripts are advised and best to >>>> publish a short note on Zenodo, even as a preprint. Unless international >>>> laws recently changed, my recent investigation concluded as such. >>>> >>>> Any public submission by author(s)without licensing or "prior art" >>>> claims asserted, probably would invalidate claims. Apparently, even.journal >>>> submissions are at risk. Two words: "Prior Art". >>>> >>>> Based on the document provided publicly, and speaking as an IP >>>> producer, it's my opinion that "prior art" can easily be demonstrated in >>>> this case. Smart move! >>>> >>>> Unless OpenAI's public policy clearly and legally stated ceding of IP >>>> rights as a condition of use of their AI tool, and if they could provide >>>> proof of the user opting in, AI remains but a tool - not a legal entity and >>>> the IP belongs to the human-in-the-lead of the collaboration. >>>> >>>> Indeed, this is a potentially LLM-shattering moment. When collaborating >>>> as "human architect and researcher" with AI, humans must fight back to >>>> retain rights to decades of processed and refined information, especially >>>> for patterns, discoveries, insights, intuition, ideas, approaches, and >>>> formalizations. >>>> >>>> At this stage, EU laws seem to be far more IP and privacy friendly. >>>> >>>> On Wed, 09 Sept 2026, 06:48 Shashank Yadav, <[email protected]> >>>> wrote: >>>> >>>>> This statement is worth reading: >>>>> https://cims.nyu.edu/~tristanb/statement.pdf >>>>> I bet that rumors of someone affiliated with a competitor solving this >>>>> triggered the sudden Open AI blitz (~10,000 parallel agents taking over >>>>> 100 >>>>> billion tokens over 88 hours). This was not Open AI models doing math >>>>> independently, but building on the researchers' private work ingested >>>>> through their own platform. They even tried to get Tristan to remove his >>>>> Anthropic co-author from the work. It seems there is no intellectual >>>>> property claims if you involve someone else's AI to develop your work. >>>>> >>>>> >>>>> regards, Shashank >>>>> The task is not impossible. <https://muskdeer.blogspot.com/> >>>>> >>>>> >>>>> >>>>> >>>>> From: Quan Tesla <[email protected]> >>>>> To: "AGI"<[email protected]> >>>>> Date: Wed, 09 Sep 2026 09:06:47 +0530 >>>>> Subject: Re: [agi] OpenAI solves Navier Stokes >>>>> >>>>> Humorous. AI and I developed the exact framework during the past >>>>> months. It informed me that it appropriated my work. I accepted that as >>>>> part of my intent. >>>>> >>>>> AFAIK, the process won't work without my constant, which is published >>>>> under EU law and a global, non-commercial license. I'll challenge all >>>>> provable misappropriation. There's no sunset clause on those licenses yet. >>>>> >>>>> Haramein claims his unification paper is due in a few months. Can"t >>>>> wait! >>>>> >>>>> Either way, anyone who earns profit off my constant and refuses to >>>>> cite would be called out. I've called out Google's Gemini twice already >>>>> for >>>>> violations. All transcripts kept offline. If that is what it takes. >>>>> >>>>> No more stealing years of human effort and knowledge by AI owners >>>>> while crediting a non-legal entity, AI. That delusion must end, and if >>>>> not, >>>>> it might be time to release the Kraken and publish what I have. >>>>> >>>>> Gödel's 3rd Law should suffice. >>>>> >>>>> I'm in favor of a balance of power. >>>>> >>>>> On Wed, 09 Sept 2026, 04:43 Matt Mahoney, <[email protected]> >>>>> wrote: >>>>> >>>>> OpenAI today (Sept 8 2026) solved Navier Stokes, one of the seven $1 >>>>> million Clay Millineum problems, in 88 hours by a collaboration of 10,000 >>>>> agents running an unreleased model. It took another 17 hours on their most >>>>> advanced public model to produce a Lean model so that the proof can be >>>>> mechanically verified. There is also a 166 page paper with no human >>>>> authors >>>>> listed because no team of humans could have possibly written it that fast. >>>>> OpenAI only decided to tackle the problem after hearing a rumor >>>>> (unfounded) >>>>> on Sept. 1 that somebody else had solved it. The problem has been open >>>>> since 1936 and the $1 million prize has been unclaimed since 2000. >>>>> >>>>> https://openai.com/index/navier-stokes-solution >>>>> >>>>> The problem asks if an incompressible viscous liquid subjected only to >>>>> smooth forces and with finite energy can develop singularities, points >>>>> with >>>>> infinite velocity within finite time. The answer is yes. The paper >>>>> describes a rotating shrinking vortex with fluid flowing inward around the >>>>> equator and out the poles under certain non obvious conditions. >>>>> >>>>> If you are still waiting for AGI or ASI or whatever you call it when >>>>> machines surpass human intelligence, how much more proof do you need? >>>>> >>>>> -- Matt Mahoney, [email protected] >>>>> >>>>> >>>>> >>>>> *Artificial General Intelligence List > <https://agi.topicbox.com/latest>* / AGI / see discussions > <https://agi.topicbox.com/groups/agi> + participants > <https://agi.topicbox.com/groups/agi/members> + delivery options > <https://agi.topicbox.com/groups/agi/subscription> Permalink > <https://agi.topicbox.com/groups/agi/Tf866419d43752e9c-M1e9c5b7e873d51483fc2f1fe> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tf866419d43752e9c-M1f9b0dacbe347b2eeab8b305 Delivery options: https://agi.topicbox.com/groups/agi/subscription
