News from October 27, according to Twilight News, October 26, Fields Prize winner, Professor of Mathematics, University of ChicagoDunsonA keynote address at last weekend's intellectuals' “Mr. Sai” forum, where he systematically described his AI Thinking about the math of the times。
DUMBLEDORE STATES THAT AI'S ADVANTAGE OVER PEOPLE IS ON THE ONE HAND THAT AI HAS UNLIMITED KNOWLEDGE, AND ON THE OTHER HAND THAT AI HAS EXTREMELY LOW TRIAL/ ERROR COSTS. BUT LET'S STICK TO IT, IN MATH, IN ANYTHINGPeople are always the subjectIA IS A TOOL THAT WE USE FOR RESEARCH。
DUMBLEDORE BELIEVES THAT IF YOU TREAT AI AS A RESEARCHER, YOU SHOULD FOLLOW ACADEMIC NORMS. IF AI PROVES A VERY IMPORTANT QUESTION, A VERY REASONABLE REQUESTIT'S THE PUBLISHED AI THOUGHT CHAINI DON'T KNOW. HOWEVER, THE REFUSAL OF AI TO PUBLISH WAS CLEARLY INCONSISTENT WITH ACADEMIC PRACTICE. “THIS AI COMPANY IS NOT WILLING TO DO THIS. OF ALL THE CERTIFICATES THEY PUBLISHED, ONLY A FEW ACTUALLY PUBLISHED THE THOUGHT CHAIN, AND MOST OF THE TIME, THEY DIDSo-called protectionDistillationFor excuses, refuses to publish the thought chain and related information.”
DUNBAR SAYS THAT FOR AIYou can't distill all of us when you're the only one who's distilling all of humanity's online literature, all of humanity's knowledgeNo one actually wants to distill you, people just want to understand the thought chain, and then you think people want to distill you, and you start saying no, that's no。
Dumbledore suggested that we could use open source mathematical modelsMake sure the math industry has its own model in its own handsTHERE'S NO NEED TO RELY ON THIS KIND OF AI COMPANY. SECOND, MODELS ARE ALSO USED IN ACCORDANCE WITH ACADEMIC NORMS, INCLUDING DATA USE, THE PUBLICATION OF THOUGHT CHAINS, TRACEABILITY, ETC。
1AI notes that on September 11th, local time, 25 Felds won as the original signatories of a joint statement by the President of the United States of AmericaArtificial Intelligence is seriously misdirected in mathematics"Serious misalignment of AI in Mathematics" warns that there is a "serious error" between AI and the Mathematics Community。
THE STATEMENT STATED THAT IN THE PAST FEW MONTHS, THE MATHEMATICAL SKILLS OF LARGE LANGUAGE MODELS HAD BEEN SIGNIFICANTLY IMPROVED AND COULD RESOLVE MAJOR OUTSTANDING ISSUES IN MANY MATHEMATICAL AREAS. HOWEVER, AI PROMOTES THE USE OF MATHEMATICS AS A BENCHMARK TO SOLVE PROBLEMS THAT ARE HARMFUL TO BOTH MATHEMATICS AND MATHEMATICS。IA'S TARGET IS SERIOUSLY OUT OF LINE WITH THE MATH COMMUNITYThese are part of a broader alignment that affects other scientific and creative occupations and society as a whole。
THE STATEMENT MENTIONS THAT THESE AI SOLUTIONS ARE OFTEN ISSUED IN HASTE, WITHOUT TIME FOR FORMAL WRITING, OR FOR SEGREGATING NEW METHODS AND IDEAS, OR CITING WORK BY OTHERS. LIKE ALL CREATIVE PROFESSIONSThis raises serious signature and copying problemsI DON'T KNOW. MOREOVER, WITHOUT MATHEMATICIANS WILLING TO TAKE CHARGE OF THEIR DEVELOPMENT AND INTEGRATION INTO THE MATHEMATICS CLASSICS, THE IDEA OF AI WILL NEVER BE FULLY REALIZED, AND THE CRITICAL HUMAN TRANSMISSION CHAIN BETWEEN MATHEMATICIANS WILL BE LOST。