Reports
Report 2026-04-17 184 sources Input Tokens: 109,992 · Output Tokens: 3,570 · Total Tokens: 114,908

AI 戰情報告 (2026-04-17)

anthropicopenaigoogleawsalibabametacerebrasmozilla claude-opus-4.7gemma-4qwen-3.6pi-0.7gpt-rosalind agentcomputer-usemcproboticsopen-weightinfrastructure

AI 戰情報告 (2026-04-17)

昨日 AI 重點新聞

Anthropic 發布 Claude Opus 4.7:視覺與編程能力的大幅躍進 ref-0, ref-3, ref-164, ref-14

Anthropic 正式推出 Claude Opus 4.7,定價維持不變但性能顯著提升。新模型在視覺能力上強化了 3 倍,可處理高達 3.75 百萬像素的高解析影像,適合處理複雜圖表與螢幕截圖。此外,其編程能力在 SWE-bench Pro 提升了 11 分,並引入全新的 xhigh 思考努力等級,讓 Claude Code 運作更加精準。

OpenAI Codex 重大更新:實現「背景操作電腦」之代理功能 ref-5, ref-18, ref-52, ref-163, ref-104

OpenAI 對其開發代理工具 Codex 進行大規模改版,推出能操作 macOS 桌面應用的新功能。該功能允許 AI 代理在背景默默執行任務,不會奪取滑鼠控制權或干擾用戶現有工作,並能透過內建的 Atlas 瀏覽器進行網頁互動,標誌著 AI 超級 App 的雛形已現。

AWS 推出 Agent Registry 並領軍 MCP 協定標準化 ref-153, ref-157, ref-7

AWS 發表 Agent Registry 預覽版,為企業提供集中式的 AI 代理與工具管理目錄。該服務支持 Model Context Protocol (MCP),這是一種由 Anthropic 創始並由 Linux 基金會託管的標準,旨在打破 AI 代理與各項數據源之間的存取壁壘,解決企業內部「代理亂象」。

Google 發布開源模型 Gemma 4 並推出 Chrome 側邊 AI 瀏覽模式 ref-54, ref-19, ref-106, ref-173

Google 釋出具備多模態能力的開源模型 Gemma 4,其 256K 上下文與強化後的代理技能引起開發者關注。同時,Chrome 瀏覽器的 AI Mode 更新後支援側邊欄視窗,讓用戶能在不切換分頁的情況下,與 AI 討論正在閱讀的網頁內容。

OpenAI 與 Cerebras 簽訂 200 億美元晶片大單以擺脫 Nvidia 依賴 ref-104, ref-181, ref-180

據傳 OpenAI 已與晶片新創 Cerebras 達成價值超過 200 億美元的伺服器晶片採購協議,並可能獲得該公司股權。此舉反映了 Sam Altman 積極垂直整合底層硬體,以降低計算成本並減少對 Nvidia 的依賴,與此同時 TSMC 的利潤也因 AI 需求激增了 58%。

機器人 AI 進化:π0.7 模型展現強大泛化能力 ref-51, ref-62

新創公司 Physical Intelligence 發表了機器人模型 π0.7,證實 AI 能引導機器人完成其在訓練集中從未見過的物理任務。這被視為「物理 AI」從特定功能轉向通用智能的重要指標。

阿里巴巴開源 Qwen3.6-35B-A3B 模型,強勢對標閉源模型 ref-1, ref-3, ref-0, ref-164

阿里巴巴推出 Qwen3.6-35B-A3B,這是一款採用 MoE (混合專家) 架構的開源權重模型。實測顯示其在 Agentic Coding 任務上能與昂貴的閉源模型(如 Opus 4.7)競爭,甚至在某些視覺 SVG 生成測試中表現更優。

Mozilla 推出自託管 AI 客戶端 Thunderbolt ref-77

Mozilla 發表開源 AI 客戶端 Thunderbolt,主打企業級的自託管 (Self-hosted) 基礎設施需求。這讓重視隱私的企業能運行自有模型而不必擔心數據流向外部雲端,直接對標 OpenAI 的企業解決方案。

人機協作 OS 與逆向工程代理工具新星:Lua 與 Kampala ref-13, ref-18

新創 Lua 獲得 490 萬歐元融資,旨在建立人機協作的操作系統;而 Kampala (YC W26) 則推出 MITM 代理技術,能將任何 App 逆向工程為 API,讓 AI 代理能繞過脆弱的瀏覽器自動化,直接操控舊有系統。

中國政府對 Meta 收購 Manus 一案展開國家安全調查 ref-128

針對 Meta 斥資 20 億美元收購 AI 代理新創 Manus 一案,中國多個機構正發起調查,擔心關鍵 AI 技術外流影響國家安全,反映出 AI 核心技術已成為地緣政治博弈的戰略資產。


市場洞察

1. 「代理治理 (Agent Governance)」正式成為雲端戰場中心 ( ) ref-157, ref-7, ref-88

延續前幾日對代理權限與監控的討論,今日 AWS 發布 Agent Registry 以及對 MCP 的重注,確立了「代理管理層」將比「模型層」更具企業溢價力。當企業內部的 AI 代理、工具與資料源 (Connectors) 多達數千個時,誰能掌握那本「黃頁目錄」與「通訊協定」,誰就掌握了企業 AI 的決策中心。

2. 「去介面化」的技術實踐:背景執行與電腦操作 ( ) ref-5, ref-52, ref-18

OpenAI Codex 的更新展示了 AI 代理不再需要模擬人類操作滑鼠,而是直接在背景 (Background) 透過作業系統層級進行任務執行。這種「去介面化」的技術成熟,預示著未來的辦公模式將從「人在 GUI 介面點選」全面轉向「人對後台代理下達 Goal-oriented 指令」。

3. 開源模型對閉源權威的「邊際成本」挑戰 ( ) ref-0, ref-3, ref-47

Qwen3.6 與 Gemma 4 的發表再次縮小了開源與閉源模型的距離。Simon Willison 的測試顯示 (ref-134),售價高昂的 Opus 4.7 在特定視覺任務上甚至輸給了能在筆電運行的開源模型。這將驅動企業在選擇模型時更加「任務導向」,並可能導致閉源廠商面臨更嚴峻的降價壓力。

4. 物理世界數據與泛化能力的匯流 ( ) ref-51, ref-62, ref-73

Manycore IPO 的轉型與 Physical Intelligence 模型 π0.7 的突破,顯示 AI 的戰場正從「數位內容生成」跨入「物理操作數據」。當 AI 代理具備泛化能力,能夠應對未經訓練的環境,商業自動化的範圍將從辦公室軟體擴展到倉儲、零售與工廠地板。

Sources 184 items
ref-0 Hacker News Claude Opus 4.7 ref-1 Hacker News Qwen3.6-35B-A3B: Agentic coding power, now open to all ref-2 Hacker News Cloudflare Email Service ref-3 Hacker News Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7 ref-4 Hacker News Cloudflare's AI Platform: an inference layer designed for agents ref-5 Hacker News Codex Hacked a Samsung TV ref-6 Hacker News AI cybersecurity is not proof of work ref-7 Hacker News Android CLI: Build Android apps 3x faster using any agent ref-8 Hacker News New unsealed records reveal Amazon's price-fixing tactics, California AG claims ref-9 Hacker News Artifacts: Versioned storage that speaks Git ref-10 Hacker News Guy builds AI driven hardware hacker arm from duct tape, old cam and CNC machine ref-11 Hacker News Show HN: MacMind – A transformer neural network in HyperCard on a 1989 Macintosh ref-12 Hacker News GPT‑Rosalind for life sciences research ref-13 Hacker News Launch HN: Kampala (YC W26) – Reverse-Engineer Apps into APIs ref-14 Hacker News Show HN: SPICE simulation → oscilloscope → verification with Claude Code ref-15 Hacker News Show HN: Marky – A lightweight Markdown viewer for agentic coding ref-16 Hacker News The beginning of scarcity in AI ref-17 Hacker News FCC exempts Netgear from ban on foreign routers, doesn't explain why ref-18 OpenAI Blog Codex for (almost) everything ref-19 Google AI Blog A new way to explore the web with AI Mode in Chrome ref-20 Google AI Blog New ways to create personalized images in the Gemini app ref-21 Google Developers Blog MaxText Expands Post-Training Capabilities: Introducing SFT and RL on Single-Host TPUs ref-22 Google Developers Blog ADK Go 1.0 Arrives! ref-23 Google Developers Blog Boost Training Goodput: How Continuous Checkpointing Optimizes Reliability in Orbax and MaxText ref-24 Google Developers Blog Build a smart financial assistant with LlamaParse and Gemini 3.1 ref-25 NVIDIA Blog No Need for Space Gear — Capcom’s ‘PRAGMATA’ Joins GeForce NOW on Launch Day ref-26 Vercel Blog How GitBook serves 30,000 sites with sub-second content updates ref-27 Vercel Blog A new programming model for durable execution ref-28 Vercel Blog Zero Data Retention on AI Gateway ref-29 Vercel Blog How FLORA shipped a creative agent on Vercel's AI stack ref-30 Vercel Blog SERHANT.'s playbook for rapid AI iteration ref-31 Vercel Blog Vercel Flags is now generally available ref-32 Vercel Blog Claude Opus 4.7 on AI Gateway ref-33 Vercel Blog Use and manage Vercel Sandbox directly from the Vercel CLI ref-34 Vercel Blog Summary of CVE-2026-23869 ref-35 Vercel Blog Vercel Sandbox now supports up to 32 vCPU + 64 GB RAM configurations ref-36 Vercel Blog Chat SDK adds Liveblocks support ref-37 Vercel Blog Transfer Marketplace resources between teams ref-38 Vercel Blog Activity Log now available in Vercel CLI ref-39 Vercel Blog View specific error codes in runtime logs ref-40 AWS ML Blog Cost-efficient custom text-to-SQL using Amazon Nova Micro and Amazon Bedrock on-demand inference ref-41 AWS ML Blog Transform retail with AWS generative AI services ref-42 AWS ML Blog How Automated Reasoning checks in Amazon Bedrock transform generative AI compliance ref-43 The Rundown AI Allbirds ditches sneakers for AI compute ref-44 Latent Space [AINews] Anthropic Claude Opus 4.7 - literally one step better than 4.6 in every dimension ref-45 Simon Willison datasette 1.0a28 ref-46 Simon Willison llm-anthropic 0.25 ref-47 Simon Willison Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7 ref-48 TechCrunch AI Factory hits $1.5B valuation to build AI coding for enterprises ref-49 TechCrunch AI Luma launches AI-powered production studio with faith-focused Wonder Project ref-50 TechCrunch AI Upscale AI in talks to raise at $2B valuation, says report ref-51 TechCrunch AI Physical Intelligence, a hot robotics startup, says its new robot brain can figure out tasks it was never taught ref-52 TechCrunch AI OpenAI takes aim at Anthropic with beefed-up Codex that gives it more power over your desktop ref-53 TechCrunch AI Anthropic CPO leaves Figma’s board after reports he will offer a competing product ref-54 TechCrunch AI Google now lets you explore the web side-by-side with AI Mode ref-55 TechCrunch AI InsightFinder raises $15M to help companies figure out where AI agents go wrong ref-56 TechCrunch AI AI traffic to US retailers rose 393% in Q1, and it’s boosting their revenue too ref-57 TechCrunch AI Roblox’s AI assistant gets new agentic tools to plan, build, and test games ref-58 TechCrunch AI Google is now targeting bad ads over bad actors ref-59 TechCrunch AI Runway CEO says AI could help Hollywood make 50 films instead of one $100M blockbuster ref-60 TechCrunch AI Meta raises Quest 3 and Quest 3S prices due to RAM shortage ref-61 TechCrunch AI Canva’s AI assistant can now call various tools to make designs for you ref-62 TechCrunch AI This simulation startup wants to be the Cursor for physical AI ref-63 The Verge AI OpenAI’s big Codex update is a direct shot at Claude Code ref-64 The Verge AI Google’s AI Mode update lets you open links without leaving the page ref-65 The Verge AI Gemini can now pull from Google Photos to generate personalized images ref-66 The Verge AI Anthropic releases a new Opus model amid Mythos Preview buzz ref-67 The Verge AI Ronan Farrow on Sam Altman’s ‘unconstrained’ relationship with the truth ref-68 The Verge AI Character.AI’s new Books mode turns reading into roleplay ref-69 The Verge AI Canva’s AI 2.0 update goes all in on prompt-powered design tools ref-70 MIT Technology Review AI Making AI operational in constrained public sector environments ref-71 MIT Technology Review AI Treating enterprise AI as an operating layer ref-72 MIT Technology Review AI Why having “humans in the loop” in an AI war is an illusion ref-73 TechMeme Physical Intelligence says its new model, π0.7, can direct robots on tasks they weren't trained on, an "early sign" of generalization, surprising researchers (Connie Loizos/TechCrunch) ref-74 TechMeme Sources: OnlyFans is in advanced talks to sell a below 20% stake at a $3B+ valuation; previous reports said it wants to sell a ~60% stake at a ~$5.5B valuation (Financial Times) ref-75 TechMeme Sources: China mobilizes multiple agencies to probe Meta's $2B Manus deal; some officials worry aggressive measures may send chilling signals to the tech sector (Financial Times) ref-76 TechMeme Hangzhou-based Manycore shares rose 187% early in its Hong Kong debut after raising $156M in its IPO; it is pivoting to selling AI training data to robot makers (Bloomberg) ref-77 TechMeme Mozilla launches Thunderbolt, an open-source AI client for users and businesses who want to run their own self-hosted AI infrastructure, available on GitHub (Kyle Orland/Ars Technica) ref-78 TechMeme Q&A: Ronan Farrow on his Sam Altman profile in The New Yorker, how Altman changed in 18 months of reporting, OpenAI's board drama and enterprise shift, and more (Nilay Patel/The Verge) ref-79 TechMeme Sources: Cerebras plans to make its IPO public as soon as Friday, aiming to raise $3B+ at a $35B+ valuation, a 60% premium to its $22B February valuation (The Information) ref-80 TechMeme Sources: OpenAI agrees to pay Cerebras $20B+ to use its server chips, double the amount previously associated with the deal, and may receive equity in Cerebras (The Information) ref-81 TechMeme Roblox agrees to pay $12.5M and implement age verification for all users as part of a settlement with Nevada over claims it failed to protect young users (Jessica Hill/Associated Press) ref-82 TechMeme Sources: SpaceX told employees that it has moved up the scheduled date for when shares awarded to employees become eligible for sale from May to April (Bloomberg) ref-83 TechMeme Alibaba unveils Qwen3.6-35B-A3B, an open-weight MoE model with 35B total and 3B active parameters, saying it rivals larger dense models in agentic coding tasks (Qwen) ref-84 TechMeme IAB and PwC: digital ad revenue hit $294.6B in 2025, up 13.9% YoY; social media advertising hit $117.7B, up 32.6%, and digital video ads hit $78B, up 25.4% (George Winslow/TV Tech) ref-85 TechMeme The US and the Philippines agree to launch a high-tech industrial hub on the island of Luzon to counter China's supply chains; the US will run it as an SEZ (Wall Street Journal) ref-86 TechMeme NYC-based Nas.com, an AI platform founded by content creator Nas Daily that helps solo entrepreneurs build online stores, raised a $27M Series A led by Khosla (Robert Scammell/Business Insider) ref-87 TechMeme Netflix says its 2026 ad revenue remains on track to reach $3B, doubling from 2025, and it now works with over 4,000 advertising clients, up 70% YoY (Bill Bradley/Adweek) ref-88 InfoQ AI/ML AWS Launches Agent Registry in Preview to Govern AI Agent Sprawl Across Enterprises ref-89 InfoQ AI/ML AWS Introduces S3 Files, Bringing File System Access to S3 Buckets ref-90 InfoQ AI/ML Google Opens Gemma 4 Under Apache 2.0 with Multimodal and Agentic Capabilities ref-91 InfoQ AI/ML Cloudflare Launches Code Mode MCP Server to Optimize Token Usage for AI Agents ref-92 InfoQ AI/ML Presentation: From VR to Flat Screens: Bridging the Input and Immersion Gap ref-93 InfoQ AI/ML Platform as a Product: Delivering Value While Balancing Competing Priorities ref-94 InfoQ AI/ML Cursor 3 Introduces Agent-First Interface, Moving Beyond the IDE Model ref-95 The New Stack As agentic AI explodes, Amazon doubles down on MCP ref-96 The New Stack Hugging Face pushes into “computer use” with HoloTab agent that works through your browser ref-97 The New Stack Who will maintain the web when PHP’s veterans retire? ref-98 The New Stack Claude Opus 4.7 arrives with better vision, memory, and instruction-following ref-99 The New Stack OpenAI’s superapp is taking shape as Codex goes beyond coding ref-100 The New Stack Dogfooding and platforms: Spotify’s agentic-first development ref-101 The New Stack Expo bets big on React Native’s agentic future ref-102 The New Stack Anthropic lays down identity verification on Claude ref-103 The New Stack Is your internal platform ready to keep up with AI-accelerated development? ref-104 Wired AI The Battle for OpenAI’s Soul ref-105 Wired AI The UK Launches Its $675 Million Sovereign AI Fund ref-106 Wired AI Google's AI Mode Update Tries to Kill Tab Hopping in Chrome ref-107 Wired AI Anthropic Plots Major London Expansion ref-108 Wired AI This Beanie Is Designed to Read Your Thoughts ref-109 TechCrunch Startups From the Startup Battlefield stage to the International Space Station: geCKo Materials built a sticky product ref-110 TechCrunch Startups Slash, a Ramp competitor founded by teenagers, raises $100M at $1.4B valuation ref-111 TechCrunch Startups You’ve heard of hybrid cars. Now meet a hybrid cement plant. ref-112 Stratechery An Interview with F1 Driver and Venture Capitalist Nico Rosberg About the Drive to Win ref-113 Tom's Hardware AMD Ryzen 9 9950X3D2 appears on Amazon with $1,000 pre-order price — AMD confirms recommended pricing is still set at $899 [Updated] ref-114 Tom's Hardware AMD's market cap hits all-time high, Intel hits 25-year high on Agentic AI's insatiable demand for CPUs ref-115 Tom's Hardware Quantum photonics roadmap — how Xanadu and PsiQuantum are looking to transfer qubits through beams of light ref-116 Tom's Hardware Meta raising Quest headset prices due to AI-driven RAM shortage — Quest 3 to cost $600, Quest 3S $350 from April 19 ref-117 Tom's Hardware Elegoo announces the Jupiter 2 resin 3D printer for $949, early bird price of $849 — new model offers massive print volume but is still physically smaller than previous models ref-118 Tom's Hardware Pragmata PC performance tested: 18 GPUs take us to the Moon ref-119 Tom's Hardware Chinese fabs import record volumes of US chipmaking equipment via Singapore and Malaysia — homegrown tool makers booked record 2025 revenues as price competition squeezes margins ref-120 Tom's Hardware Get an entire Lenovo Legion Pro 7i RTX 5090 gaming laptop with OLED screen for $3,199, less than the price of a desktop RTX 5090 graphics card — save $800 while stocks last on this monster machine ref-121 Tom's Hardware US appeals court restarts $3 billion patent infringement lawsuit against Intel — VLSI case from 2017 returns after court sets aside 2024 decision ref-122 Tom's Hardware Two US citizens get combined 16 years in prison for running North Korean laptop farms — fake remote IT work scheme netted DPRK $5 million in around three years ref-123 Tom's Hardware Intel launches Wildcat Lake as Core Series 3 for value laptops and edge systems — six consumer SKUs built on 18A promise 'all-day' battery life ref-124 Tom's Hardware Legendary Qualcomm, Apple, and Nuvia alumni form new CPU startup — Nuvacore promises to 'rewrite the rules of silicon' ref-125 Tom's Hardware Bambu updates its 3D printers to print unique hues or gradients using two or three filaments — company acknowledges OrcaSlicer-FullSpectrum fork as the basis for the color prediction part of the new feature ref-126 Tom's Hardware Score 64GB of G.Skill DDR5 and a 1TB 990 Pro for an insane $441 in this Newegg bundle — 9800X3D and X870E motherboard also comes with a free AIO, just $1,299 to start an AM5 gaming build ref-127 Tom's Hardware Asus TUF Gaming A14 (2026) review: Big CPU power in a mispriced gaming laptop ref-128 Tom's Hardware Broadcom to supply Meta with custom silicon through 2029 — Broadom CEO Hock Tan departs Meta's board ref-129 Tom's Hardware Anonymous perps behind 86 million files scraped from Spotify hit with $322 million court judgement — Anna's Archive case presents intriguing precedent for AI training ref-130 Tom's Hardware Non-functioning counterfeit Samsung 990 Pro SSDs are circulating in Europe — Despite convincing packaging, blue PCB easily gives away the dupe ref-131 Tom's Hardware Oklahoma farmer arrested and jailed for trespassing during AI data center town hall — removed by officers after going a few seconds over allotted speaking time, trying to hand paperwork to counselors ref-132 Tom's Hardware Virginia voter support for new data centers collapses from 69% in 2023 to 35% in new poll — Multi-gigawatt, 37-building Digital Gateway project abandoned ref-133 Tom's Hardware Struggling shoemaker and apparel brand Albird pivots to AI data centers, stock jumps 580% in a single day — sells core business and leveraging $50 million in financing to become a GPU-as-a-Service and AI cloud solutions provider ref-134 Tom's Hardware IPv6 usage reaches historic 50% across Google services, matching IPv4 — increased usage eases pressure on the IPv4 address market as 'new' protocol designed in 1998 finally hits its stride ref-135 Tom's Hardware Engineer open-sources DIY radar system that's 95% cheaper than $250,000 commercial offerings, has 20 kilometer range — Moroccan engineer designs Aeris-10 radar, shares it on GitHub ref-136 Politico Tech GOP finalizing draft national privacy law that would preempt states ref-137 The Hacker News (Security) NIST Limits CVE Enrichment After 263% Surge in Vulnerability Submissions ref-138 The Hacker News (Security) Operation PowerOFF Seizes 53 DDoS Domains, Exposes 3 Million Criminal Accounts ref-139 The Hacker News (Security) Apache ActiveMQ CVE-2026-34197 Added to CISA KEV Amid Active Exploitation ref-140 The Hacker News (Security) Newly Discovered PowMix Botnet Hits Czech Workers Using Randomized C2 Traffic ref-141 The Hacker News (Security) ThreatsDay Bulletin: Defender 0-Day, SonicWall Brute-Force, 17-Year-Old Excel RCE and 15 More Stories ref-142 The Hacker News (Security) [Webinar] Find and Eliminate Orphaned Non-Human Identities in Your Environment ref-143 The Hacker News (Security) Cisco Patches Four Critical Identity Services, Webex Flaws Enabling Code Execution ref-144 The Hacker News (Security) Hidden Passenger? How Taboola Routes Logged-In Banking Sessions to Temu ref-145 The Hacker News (Security) Obsidian Plugin Abuse Delivers PHANTOMPULSE RAT in Targeted Finance, Crypto Attacks ref-146 EU Startups Swedish semiconductor startup AlixLabs closes €15 million Series A to scale atomic-level etching technology ref-147 EU Startups Stockholm’s BioLamina secures €20 million EIB loan to scale cell therapies for chronic diseases ref-148 EU Startups London-based startup Lua raises €4.9 million to scale its human-agent collaboration platform ref-149 EU Startups As the AI era strains human verification, Danish startup Flare bags €3.6 million to support knowledge validation ref-150 EU Startups STORM Therapeutics raises €47.5 million Series C to harness the power of RNA modification to treat cancer ref-151 EU Startups The EU-Startups Podcast | Interview with Patrick Andrae, Co-founder and CEO of HomeToGo ref-152 EU Startups ALL THINGS raises €4.1 million after chef Thomas Straker’s viral butter series sparked a dairy brand ref-153 Lobsters Reimplementing the Space Protocol Stack from Scratch ref-154 Google News Finance ministers and top bankers raise serious concerns about Mythos AI model ref-155 Google News Factory hits $1.5B valuation to build AI coding for enterprises ref-156 Google News LinkedIn’s AI Agent Product Becomes a Surprise Bright Spot for Microsoft ref-157 Google News AWS Launches Agent Registry in Preview to Govern AI Agent Sprawl Across Enterprises ref-158 Google News OpenAI launches AI model GPT-Rosalind for life sciences research ref-159 Google News How Automated Reasoning checks in Amazon Bedrock transform generative AI compliance ref-160 Google News AI is finally delivering productivity — for remote employees ref-161 Google News Meta debuts first AI model from new superintelligence group ref-162 Google News Apple sends Siri developers to AI coding bootcamp ref-163 Google News OpenAI Brings Major Codex Update: AI Agent 'Can Now Operate Your Computer' ref-164 Google News Anthropic Unveils Opus 4.7: A Leap in AI Coding and Vision ref-165 Google News White House grants US agencies access to Anthropic’s Claude Mythos AI model ref-166 Google News How Alibaba’s new AI model is challenging Tencent and reshaping the global tech race ref-167 Google News Perplexity Turns Mac mini Into Local AI Agent ref-168 Google News New ILO–World Bank paper highlights uneven global impact of generative AI on jobs ref-169 Google News Google, CoreWeave Fuel AI Funding Frenzy With $6.7 Billion Bonds ref-170 Google News AI startup offers humans ‘perfect and infinite memory’, Harvard professor says ref-171 Google News Stellantis and Microsoft sign 5-year AI partnership ref-172 Google News Accel raises $5 billion fund to scale late-stage AI startup investments ref-173 Google News Google and Pentagon: A Strategic AI Partnership in the Making ref-174 Google News Anthropic Receives Investment Offers Valuing AI Startup at Up to $800 Billion ref-175 Google News Salesforce, Databricks unveil agentic AI governance tools ref-176 Google News Russian businesses warn AI regulation bill would raise costs and restrict access to advanced technology ref-177 Google News Anthropic safety chief quits with dire warning ref-178 Google News Govt forms high-level inter-ministerial body to steer AI governance strategy ref-179 Google News Allbirds drops sneakers, reinvents itself as an AI infrastructure company ref-180 Google News OpenAI signs massive AI chip deal with Cerebras ref-181 Google News TSMC Posts Record Q1 Profit, Up 58% on AI Chip Demand ref-182 Google News Broadcom (AVGO) Is Up 13.1% After Expanded Meta AI Chip Partnership - What's Changed ref-183 Google News Korean Professors Pioneered GPU Use for AI Before NVIDIA