X

Holy, rumors are spreading everywhere that OpenAI is close to verifying a proof of the Hodge conjecture, while either OpenAI or Anthropic may be nearing a solution to Birch–Swinnerton-Dyer. Both are Millennium Prize Problems that have resisted decades of mathematical research. Verified AI-generated proofs of both would be a historic achievement for mathematics and a remarkable demonstration of AI’s ability to produce new scientific knowledge. And infact show that AI is capable of finding novel and creative solutions.

太丝滑了朋友们！还要啥自行车！ 我用 GPT-6 Astra 1.5h 完美复刻 Apple 的空间 3D 景深效果，正好适配 iPhone Duo 折叠屏丝滑交互/动画！ GPT-6 我还要给你充钱！！！ https://t.co/kp2yQlTOQ2

这不就成了嘛朋友们！人物动起来 + 翻折真神了！ 用 GPT-6 Astra 实现 Apple 的任意图片 3D 景深效果，完美适配 iPhone Duo 的丝滑交互！ https://t.co/Boe4hiHGYR https://t.co/AtctrRHINW

DeepSeek just released V4.1-Flash with a new architecture, six weeks after its July V4-Flash update. July’s release improved post-training while keeping the architecture unchanged. (Same with GLM-5.3/Flash) V4.1 introduces a Causal Encoder–Decoder architecture with native visual understanding. DeepSeek reports: - 552B MoE parameters, with 8B active during input processing and 16B during output generation. - KV-cache requirements cut to ¼ of the HBM and ⅛ of the SSD storage versus the previous generation. - Lower API prices. These are *significant* jumps in just a few weeks with post training. This is the new reality we have to adapt to: weekly releases with significant improvements. The company says Flash now beats V4-Pro on capability, cost and speed. Starting September 14, V4-Pro API requests will temporarily route to V4.1-Flash until V4.1-Pro arrives.

赶紧用 GPT-6 Astra 玩起来吧！iPhone Duo 交互与动画纵享丝滑！ （还在让 GPT-6 做把普通图片变成官网多层 3D 景深的开发，才能完美适配 https://t.co/Xgq2qKF9L9 https://t.co/Fe3KYh65Nf

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest model in our new architecture family, with native visual understanding. 🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models. 1/6

DAILY AI BRIEF 🗞 — Sept 9 META 🔥: > Muse is live. Internal name was Hatch. US-only personal agent on iOS, Android, web, and WhatsApp. > Free tier plus $20 and $100 monthly plans. Powered by Muse Spark 1.3. Smart glasses later! OPENAI 🔥: > ChatGPT Images 2.5 is out for all ChatGPT, Work, and Codex users. Up to 50% faster, sharper detail, more reliable multi-turn edits. > New Sketch tool in ChatGPT. API adds GPT-Image-2.5 Flare (fast default) and Sunburst (higher control). > An unreleased next-gen model, run with ~10,000 agents for 88 hours, produced a claimed solution to the Navier–Stokes Millennium Prize Problem. Independent verification still pending. ANTHROPIC 🔥: > Pretraining researcher Jacob Coxon resigned. Says neither OpenAI nor Anthropic is acting responsibly, and both are racing to self-improving superintelligence. > Alignment Science lead Evan Hubinger: “We really do earnestly believe AI could kill all humans.” His personal odds: >10% this decade. No plan yet for superintelligence alignment. X 🔥: > Original Content Rewards is rolling out. Old Creator Revenue Sharing ended Sept 7. > Last old payout around Sept 11. First new payout Sept 25 if you get approved. I wasn't :P GOOGLE 🔥: > Google DeepMind shipped AlphaGenome Atlas — predicted effects for all 9 billion single-letter human DNA variants, plus a single AVI impact score. Free for non-commercial use. * Used Grok to compose this brief, cherry-picking the news and doing some post-editing.

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest model in our new architecture family, with native visual understanding. 🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models. 1/6