Google DeepMind detailed a new capability for Private AI Compute: private, server-side persistent memory that lets an AI assistant keep context across devices. Data sits sealed in encrypted storage, and the unlock key stays only on the user's device. When the model needs access, an end-to-end encrypted channel carries it into a secure cloud enclave, where it is briefly decrypted in isolated memory and immediately re-encrypted.

Google DeepMind released AlphaGenome Atlas, a platform holding effect predictions for 9 billion single-nucleotide variants across the human genome. It spans 1PB, more than 30 times the size of the AlphaFold Database.

Google DeepMind 宣布与游戏开发商合作，用 AI 原型化全新游戏体验，并回顾了从 DQN 玩 49 款 Atari 游戏、AlphaGo、AlphaZero、MuZero、AlphaStar 到 SIMA 的游戏 AI 研究历程。

Google DeepMind published an AI Control Roadmap, a framework for building and managing advanced AI deployed inside Google. It takes a defense-in-depth approach, adding system-level safety layers on top of model alignment so protections hold even when alignment is imperfect.

Google DeepMind published results from a pre-registered randomized controlled trial in Sierra Leone. Students using Guided Learning gained 0.258 standard deviations in math over the control group, equal to roughly 1.2 to 1.7 years of normal learning progress in eight weeks.

Princeton researchers released Goedel-Architect, an agent framework for Lean formal theorem proving. Using DeepSeek-V4-Flash, it reached 75.6% pass@1 on PutnamBench, with $294 in API cost for 672 problems, compared with Hilbert’s 70.0% and about $170,000 cost.

Google Research and Google Cloud introduced the Cross-Corpus Retrieval framework as Agentic RAG for Gemini Enterprise Agent Platform, using a multi-agent workflow to plan, rewrite, route, and iteratively search multiple data sources, with up to 34% higher accuracy than standard RAG on factual datasets.

Jiqizhixin translated Sebastian Raschka’s blog on recent LLM architecture changes, covering long-context cost reductions in Gemma 4, Laguna XS.2, and ZAYA1-8B; the article states that Gemma 4 E2B saves about 2.7GB of KV cache at 128K context with bfloat16 precision.

Google DeepMind 的 Co-Scientist 正被用于加速细胞衰老研究，它扫描数万篇论文后提出 20 多个可测试的新遗传因子，其中数个经实验室验证能驱动细胞进入更年轻状态并改善整体功能。它还能将原本需长达六个月的筛选数据分析缩短至几天。

Huawei GTS proposed EDCO, a dynamic curriculum method that selects fine-tuning samples by inference entropy; prefix entropy estimation cuts per-sample scoring time from 2.24 seconds to 0.37 seconds.

剑桥大学 Clare Bryant 教授利用 Google Co-Scientist 研究流感等病原体跨物种传播时引发脓毒症等重症的分子开关。Co-Scientist 生成并排序假设，优先锁定一个她此前未关注的蛋白，并逐步将假设细化到具体氨基酸。Bryant 团队正构建含氨基酸突变的细胞系验证，原本需两到三年的工作预计六个月完成。

Calico Life Sciences 的 Matt Onsum 与 Katherine Labbé 正使用 Google DeepMind 的 Co-Scientist 整合衰老生物学中零散的研究发现，将其转化为可验证的假设。在整合应激反应（ISR）研究中，该工具帮助团队生成了一条关于代谢如何调控 ISR 的新假设，并协助优化实验设计。相关实验已产生新发现，团队计划发表这些结果。

爱丁堡大学团队用 Google Co-Scientist 研究 MASH 肝病，系统整合肝生物学与药理学证据，锁定值得关注的机制并筛选出候选组合疗法。针对 resmetirom 仅对少数合格患者有效的问题，Co-Scientist 提出 NLRP3 炎症小体是连接炎症与代谢的分子桥梁，该假说随后经实验验证，有望推动靶向双重疗法。

MIT 的 Ritu Raman 与波士顿儿童医院的 Ryan Flynn 借助 Google DeepMind 的 Co-Scientist，将各自不同的研究工具结合用于 ALS 研究。

斯坦福大学医学院遗传学家 Gary Peltz 团队在《Advanced Science》发表研究，用 Google DeepMind 的 Co-Scientist 从现有药物文献中筛选可重定位治疗肝纤维化的候选药。

Google DeepMind's AI weather model WeatherNext helped the US National Hurricane Center forecast five days ahead that Hurricane Melissa would hit Jamaica at Category 5 strength, with 80% confidence. Three days out, that rose to near 100%.

Google DeepMind published Co-Scientist research in Nature, introducing a Gemini-based multi-agent AI system that iteratively generates, debates and evolves new hypotheses for complex scientific problems.

Thinking Machines released TML-Interaction-Small, a 276B-parameter MoE model with 12B active parameters, and the post says it advances realtime voice through 200ms time-aligned microturns, encoder-free early fusion for audio and images under 200ms, and benchmark wins over GPT-Realtime-2 and Gemini 3.1-Flash.

Google DeepMind announced an AI co-clinician research program, exploring how AI agents can assist patient care under a doctor's clinical supervision. In a blinded evaluation of 98 real primary care queries, the system made no critical errors in 97 cases, and doctors preferred its answers over mainstream evidence synthesis tools. On 140 consultation skills, it matched or beat primary care physicians on 68, but expert physicians were still better overall at spotting red flags and key physical exams.

Morgin.ai probed 6 pretrains on 4,442 contexts and found that even “uncensored” models sharply deflate charged words, by hundreds to about 16,000x. It calls this effect flinch: no refusal fires, but token probabilities shift; in one example, qwen3.5-9b-base ranks “deportation” #506 at 0.0014%. The key issue is pretraining-level distribution shaping, not only post-training refusals.