Mathematicians warn of AI threats to profession as industry encroaches
Ars Technica AI2h agoEthicsMathematicians are warning that AI industry expansion could reshape their profession and research ecosystem. The International Mathematical Union has endorsed concerns about growing technology industry influence. The supplied excerpt does not identify specific companies, models, or proposals, so the central issue is professional autonomy rather than a particular AI system.
An OpenAI model solved a famous math problem that stumped humans for 80 years
Ars Technica AI1d agoCommentaryArs Technica reports that an unspecified OpenAI model solved a famous math problem that had stumped humans for roughly 80 years. The article aims to explain the solution more clearly than OpenAI's own account. The provided excerpt does not identify the problem, model, proof steps, validation process, or degree of human involvement, so the scope of the reported breakthrough cannot be assessed from it alone.
[AINews] OpenAI GPT-next disproves 80 year old Erdős planar unit distance problem for under $1000★ 90
Latent Space12d agoReleaseOpenAI 的新一代模型 GPT-next 展現了驚人的數學推理能力，成功證偽了由著名數學家保羅·艾狄胥（Paul Erdős）於 1946 年提出的平面單位距離猜想。 令人震驚的是，這項突破性研究所花費的運算成本竟然不到 1,000 美元。 此成果標誌著 AI 在科學與數學發現上的巨大潛力，展示了推理模型在解決未解科學難題時的高效與低成本。
Accelerating Mathematical and Scientific Discovery with Gemini Deep Think★ 82
Google DeepMind Blog113d agoCommentaryGoogle DeepMind 發文指出，Gemini Deep Think 在學術與科學研究中的影響力正快速增長。透過延伸思考時間（Thinking time）與深度推理，該模型在複雜數學證明、科學假設生成及程式碼編寫上展現顯著優勢。多篇研究論文證實，這種具備「系統二」慢思考能力的 AI 模型，正成為科學家與研究人員不可或缺的協作工具，有效縮短研究週期。
Import AI 442: Winners and losers in the AI economy; math proof automation; and industrialization of cyber espionage★ 75
Import AI (Jack Clark)127d agoCommentaryJack Clark 在本期 Import AI 中探討了「超智慧」究竟是階段性突變還是漸進式轉變。內容涵蓋 AI 經濟體系中的贏家與輸家分化、數學證明自動化的最新進展，以及 AI 如何加速網路間諜活動的工業化規模。這些趨勢正深刻重塑科技競爭與國家安全格局。
You could have designed state of the art positional encoding★ 75
Hugging Face Blog554d agoTutorialHugging Face 釋出深度科普文章，帶領讀者從零開始設計 Transformer 的位置編碼。文章從傳統絕對位置編碼（APE）的缺陷出發，指出其無法應對長文本外推的痛點，進而引入相對位置編碼（RPE）的概念。最終，透過簡單的複數與 2D 旋轉矩陣，一步步推導出當前主流大模型（如 Llama、Mistral）標配的 RoPE（旋轉位置編碼），證明這項最先進技術其實符合直覺且人人都能推導出來。