与「NOA」相关的搜索结果

I asked people: "who are the best active angels in AI?" 5 votes: Jeff Dean (@JeffDean) 4 votes: Umesh Khanna (@forwarddeploy) 3 votes: Sholto Douglas (@_sholtodouglas) 3 votes: Thomas Wolf (@Thom_Wolf) 2 votes: Daniel Chang (@dnnssl2) 2 votes: Elad Gil (@eladgil) 2 votes: Eric Glyman (@eglyman) 2 votes: Logan Kilpatrick (@OfficialLoganK) 2 votes: Michael Chen (@michaelzchen5) 2 votes: Nat Friedman (@natfriedman) 2 votes: Noam Brown (@polynoamial) 2 votes: Philip Bogdanov (@philip_bogdanov) 2 votes: Sarah Zeng (@sarahzengy) 2 votes: Tim Luong (@chongz) 2 votes: Zach Frankel I asked people to focus on angels that work in AI. People couldn't pick themselves. 3 people picked me; I left myself off the list. I didn't vote.

据彭博社，Binance 从 CryptoCom 聘请了两名高级合规主管，以加强其合规团队。Antonio Alvarez 将担任 Binance 首席合规官 Noah Perlman 的副手，Duncan DeVille 将出任全球金融犯罪合规负责人。两人此前均在 CryptoCom 任职，分别拥有 Coinbase、Visa、Western Union 及美国财政部等机构的合规经验。此次人事调整发生在 Binance 过去一年多名合规人员离职以及持续面临监管审查的背景下。2023 年，Binance 因未维持有效反洗钱合规体系，与美国司法部门达成 43 亿美元和解协议，此后持续强化合规体系。

谷歌组织架构大洗牌！ Demis Hassabis 卸任 Google Deepmind 的 CEO，转任 GDM 主席兼 Alphabet 首席科学家，同时继续领导 Isomorphic Labs，把主要精力放在 AGI 的长远方向、全球战略以及用 AI 推动科学与医疗突破上； “我们已抵达人类历史上的一个关键时刻。我一生都在为 AGI 努力，现在，就像你们许多人一样，我感觉它近在咫尺” —— Demis Hassabis Koray Kavukcuoglu 从 Deepmind CTO、Google 首席 AI 架构师升任 Google DeepMind 高级副总裁，直接向 Pichai 汇报，全面负责 Gemini 模型研发、前沿 AI 研究以及 Gemini 应用与开发者团队。其实就是接替 Demis 的管理工作，他在 DeepMind 已工作 13 年，，曾主导 WaveNet、DQN 等成果； 在谷歌工作 27 年的 Jeff Dean 将与资深研究员 Sanjay Ghemawat 一起离开，创办一家独立的公益性质公司 Discovery Loop，专注“全自动科学发现的闭环”研究；Google 会作为创始投资方和云服务合作伙伴继续与他们合作。Dean 告诉 NYT，离开 Google 让他有更多余地专注于科学发现； 三位最初的 Gemini 负责人如今全部离职：Noam Shazeer 去了 OpenAI，Dean 和 Vinyals 去了 Discovery Loop。 我如何看待这次调整？ 在 2023 年 Google 合并 Google Brain 与 Deepmind，于是有了 Gemini 的成功；三年后 Google 再次面临大公司病、组织障碍（人才保留上的结构性问题）和工程低效的挑战，于是有了这次必要又意料之外的调整。 部分传言 Demis 因 Gemini 执行不力而被边缘化，换上Koray 是希望把“研究强、产品化慢”这个 Google 的老毛病治好，对比 OpenAI / Anthropic 快速产品化，Google 需要证明也能更快的迭代。所以 Google 全村的希望都在 Gemini 4，内部确认已经有了重大进展。我保留短期观察，长期看好的判断👀 其实 Google 已经证明了目前不用靠模型的能力，只用买算力就能保持盈利优势，Google 是目前唯一的全栈 AI 科技巨头。现在决定内部加速产品化，再利用投资加孵化，既能解决人才保留的问题，又可以抢占 AGI 与科学前沿。 这是 Google 的角色优化，并非危机。执行力与人才保留正在成为新的竞争护城河，技术不再是 Google 的短板，专注和敏捷才是。

谷歌最近出啥问题了？接连损失大将。刚刚杰夫.迪恩宣布从谷歌离职创立Discovery Loop。这位非常不简单，杰夫.迪恩是计算机科学与人工智能领域的传奇人物，被誉为“谷歌的系统架构总设计师”和硅谷“程序员中的程序员”。 1、杰夫离职前担任谷歌首席科学家兼任 Google DeepMind 和 Google Research 的核心技术领袖。是Google Senior Fellow/Level 11，在谷歌数十万员工中，能达到 Level 11 的极度罕见，几乎代表了整个谷歌技术体系的终极权威。 2、Jeff Dean在谷歌工作27年，其职业生涯贯穿了互联网海量数据处理时代与现代人工智能时代，其核心成就可分为三大阶段： 1）奠定现代互联网基础的基础设施（2000年代） 在谷歌成立早期，他主导开发了一系列支撑起整个 Google 生态（乃至全行业大数据架构）的基础系统： MapReduce：划时代的分布式计算框架。它将庞大的数据处理任务拆分到数万台普通服务器上并行运行，是后来的 Hadoop 等大数据技术的直接灵感来源。 Bigtable：高扩展性、高性能的分布式结构化数据存储系统，直接催生了后来的 NoSQL 数据库浪潮（如 HBase）。 Spanner：全球级分布式 SQL 数据库，打破了传统数据库“一致性”与“可用性”难以兼得的困境。 搜索与广告系统架构：主导设计了谷歌早期的五代搜索索引与抓取系统，以及 AdSense 广告投放系统的早期基础设施。 2）推动 AI 革命与深度学习生态（2010年代） 创建 Google Brain（谷歌大脑）：2011 年与吴恩达（Andrew Ng）等人共同创办 Google Brain，将深度学习引入谷歌的核心业务。 TensorFlow：主导打造并开源了全球最著名的深度学习框架之一 TensorFlow，极大地降低了全球研究者和开发者使用 AI 的门槛。 DistBelief & Pathways：先后设计了 DistBelief（早期大规模分布式神经网络训练系统） 和 Pathways 架构（为后续训练 PaLM、Gemini 等超大规模多模态大模型奠定了系统支撑）。 3）整合与领军时代（2020年代） 统领 Google AI 与 DeepMind 融合，杰夫曾长期担任 Google AI（Google Research）负责人。2023 年 Google Brain 与 DeepMind 合并为 Google DeepMind 后，他作为首席科学家，重点统筹下一代超大模型（如 Gemini 系列）的系统架构与前沿 AI 研发。 前段时间谷歌还有两位大拿离开： 1）Google DeepMind 副总裁 John Jumper跳槽至 Anthropic，这位因带领团队打造预测蛋白质结构的革命性 AI 模型 AlphaFold，与 Demis Hassabis 共同获得 2024 年诺贝尔化学奖。 2）谷歌工程副总裁兼旗舰模型 Gemini 联合负责人Noam Shazeer跳槽至 OpenAI。 Noam Shazeer 是Transformer 架构奠基人之一：2017 年奠定现代大模型基础的划时代论文《Attention Is All You Need》的共同作者。 前线厮杀正浓，统兵大将接连离开，对谷歌影响应该不小。

ICYMI: our July AI recap ⬇️ 🌦️ Advanced weather forecasting with @NOAA and Google Cloud, using high-performance H4D virtual machines to run its atmospheric models, providing meteorologists with the AI capabilities needed to deliver life-saving early warnings. 💻 Introduced 3 new Gemini models — Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber — to deliver the efficiency, latency, and reliability to build AI agents at scale. 🇺🇸 Launched the Alliance for America’s Skilled Trades With BlackRock, Carhartt, and Ford to build a stronger pipeline for the skilled trades, scale evidence-based training approaches, and expand opportunities nationwide. ☁️ Rolled out AlphaEvolve, our Gemini-powered AI code-optimization agent, to help Cloud customers solve their hardest problems by automatically searching for better solutions and returning human-readable, optimized code. 🎵 Delivered significant advancements in musicality, lyrics, and vocal quality in Google Flow Music with Lyria 3.5, our newest music generation model. 📡 Launched three new FireSat satellites, expanding a global initiative led by the Earth Fire Alliance (EFA) with @GoogleResearch to help detect wildfires before they spread. 🧪 Established over 15 global bioresilience partnerships across governments, research groups, and biosecurity organizations to prevent model misuse, detect outbreaks rapidly, and deliver swift, coordinated responses in partnership with @IsomorphicLabs.

This week sovereign wealth-backed capital came onchain, retail giants opened Solana to 8M+ households, and tokenized asset volume smashed another all-time high. Here's everything that shipped: 📰 Headline News - Morgan Stanley made Solana available to eligible E*TRADE clients, opening access across 8.7 million households - @Mubadala Capital’s Alternative Solutions Fund went live on Solana via @KAIO_xyz - @tryramp opened 24/7 stablecoin accounts and payments settling on Solana 📰 Launches - Intel tokenized stock ($INTC) went live via @Backpack Securities and @sunrise - @jtx_trade opened its unified trading surface for majors, memes, and RWAs to everyone - @LayerZero_Core and @KeetaNetwork brought tokenized bank deposits to Solana - @MetaMask activated gasless Solana swaps, covering gas fees for trades over $200 - @Lexur launched in public beta, aggregating Solana perps across Pacifica, Jupiter, Bullet, and more - @dopplerprotocol went live on Solana Mainnet - @hylo_so shipped xBTC, bringing liquidation-resistant 3x BTC exposure onchain - @Bulletxyz launched 24/7 trading with up to 10x leverage for metals, oil, and chipmakers - @trydapital debuted Dapital, a social trading network for RWAs, memes, and tokens - @altitude introduced Altitude Accounting to automate transaction classification - @Thea_AI launched AI inference routing and settlement on Solana - @RyderWallet added native SOL staking on Ryder One via @Meria_Finance and @yieldxyz - @CandyDigital enabled true self-custody for digital collectibles on Solana - @meleemarkets unveiled Parimutuel Market Maker (PMM) design for permissionless prediction markets - @Pumpfun introduced BOOST mode to reinject post-migration liquidity into token buybacks and burns - @risedotrich enabled cross-chain swaps to Solana powered by @deloraprotocol - @Beezie went live with zero-recycled-slab claw machine collectibles - @BananaZoneApp shipped Paid Play for @solanamobile users with real rewards - @streamflow_fi enabled NFT locking for time-bound collectible custody - @backyard_fi kicked off Backyard BBQ: Season 0 for yield rewards - @cleopetrafun launched Thesis Index to trade macro theses via prediction market baskets - @flint_trade_ launched streamlined liquidity quoting for Solana market makers - @tapmapfun launched Treasure Hunt for local business check-ins and rewards - @machinedotfun, @codecopenflow, and @peaq opened applications for The Foundry, a Physical AI accelerator - @TryNoahAI opened waitlists for Noah Accelerate to support Solana builders - @magicblock and @SuperteamMY opened applications for their VIP builder program - @fairdotclub partnered with @LaFamilia_so to expand private capital formation 📰 Milestones - Tokenized-asset trading volume on Solana reached a record $5.8B in Q2, up 114% QoQ - @byreal_io surpassed $4B in total trading volume - @kamino’s OnRe Market passed $200M in total market size - @humafinance PST crossed $200M market cap - @ExponentFinance crossed $120M TVL - @usdgo_official crossed $1B in circulating supply in 5 months - @Beezie reached $2M in total volume across 5965 claw pulls - Solana consumer card top-ups reached a record $94.32M in a single month in Q2 - @fomo logged 17.4K active daily traders executing on Solana If you enjoyed this week’s newsletter, please share it with an RT. Artwork by @Suffoca 🔥

On July 14, the Department of the Navy released its Strategy to Weaponize Data and AI: a roadmap to an AI-first Fleet that can "out-learn and out-fight any adversary." Within a week, our team answered with working code. The Navy's new strategy frames data and AI as warfighting assets on par with weapons and munitions, and it prizes one thing above all: turning information into decisions, fast. That is exactly the problem our GURU architecture was built for. So when two Navy SBIR topics called for AI-driven maritime tracking and adaptive sensor management, we pointed at the sea what we had already proven in orbit. The video below shows both prototypes, back to back. First, GURU MarineGuard: 787 real vessels from public NOAA data, replayed through a cascade of learned models that forecast each ship's movement, flag deviations from its learned pattern of life, and hand analysts a ranked review queue instead of an unfiltered flood. The full stack runs on a laptop. Second, an adaptive sensor resource manager add-on to MarineGuard: it measures each radar's marginal contribution to each track, projects the consequence of releasing a sensor task before proposing it, and then waits for the operator. Advisory by design. Both inherit their DNA from OrbitGuard, our system watching 14,710 space objects at the SDA TAP Lab with 94 to 96 percent maneuver-detection accuracy. Same architecture, new domain, days not years. These are prototypes, and we say so on screen. The trajectories are real. The sensor numbers are deliberately notional. No score is a threat call. Showing your assumptions is a capability, not a caveat, and national-security guidance now demands exactly that: AI that is reliable, robust, steerable, and controllable under rigorous test and evaluation. Our doctrine was written for that bar. Learned models accelerate and rank. Validated references confirm and decide. Humans stay in command. One architecture. Space, maritime, autonomous engineering, regulated nuclear autonomy. This is simply the latest sign of what this team fields, fast, where mistakes are not allowed. Sailors, engineers, program folks: what mission should GURU learn next? #DefenseTech# #ArtificialIntelligence# #MaritimeDomainAwareness#