Loading...
Loading...
The AI race is shifting from single-request models to persistent, autonomous agents, and Anthropic’s Claude sits at the heart of that shift. Google’s I/O pushed an agent-first stack—Gemini 3.5, Antigravity 2.0, and managed server-side agents—consolidating tooling but raising control and privacy concerns. Anthropic counters with rapid product updates (Claude Code, plugins, Slack bridges), big hires like Andrej Karpathy, and massive compute deals (SpaceX/Colossus), even as its profitability claims draw scrutiny. The contest now hinges on managed execution layers, compute supply, developer ecosystems, and talent—reshaping geopolitics, cloud strategy, and who controls autonomous AI in production.
Anthropic and Google are racing to own agentic AI stacks that combine models, orchestration, and developer tooling, reshaping how teams build automated workflows. Tech professionals must weigh trade-offs between managed execution, developer control, and tooling lock-in when choosing platforms.
Dossier last updated: 2026-05-21 10:23:00
Google I/O 2026’s biggest shift wasn’t faster models but a move to persistent, server-side agent runtimes: agents that run continuously in managed environments instead of responding to single requests. Google layered Gemini 3.5 variants, Antigravity 2.0, Gemini API Managed Agents, and consumer features like Gemini Spark to place models inside durable execution contexts with scheduled tasks, subagent spawning, and cloud-hosted personal agents. A key signal: Google is deprecating some local CLI and free-tier tools, steering developers toward metered, enterprise-managed infrastructure. That raises new operational, security, and privacy questions because agents can act autonomously “while nobody is watching,” changing how developers think about boundaries, observability, and control.
Ben Thompson argues the biggest near-term shift in AI is from human-in-the-loop “answer inference” to autonomous “agentic inference,” which will reshape compute architectures and favor players beyond Nvidia, including China and space-based compute. Coverage highlights Anthropic’s surprising compute deal with xAI/Elon Musk and what it means for markets, SpaceX’s role in serving other AI labs, and the emergence of deployment-focused companies—including OpenAI’s new entity—to implement AI at scale. The newsletter also reviews US-China tech relations around a presidential visit, and commentary on Musk’s lawsuit with OpenAI. These pieces matter because they map how compute, deployment strategy, geopolitics, and corporate alliances will determine who wins in the next phase of AI.
Stratechery’s weekly roundup highlights a shift toward “agentic inference,” where AI agents operate without human involvement, reshaping compute architecture and favoring different infrastructure choices. Ben Thompson argues this will create distinct trade-offs from current training/inference paradigms and could advantage China and space-based compute. Coverage also examines Anthropic’s compute deal with xAI/SpaceX, its market logic, and implications for Musk’s strategy and space data centers. Additional pieces discuss OpenAI creating a deployment-focused company, Apple-Intel ties, and commentary on Elon Musk’s lawsuit against OpenAI. The newsletter frames these developments as central to how AI deployment, corporate alliances, and geopolitical dynamics will influence tech infrastructure and market power.
断供OpenAI，Anthropic买下全球1/4开发者都在用的工具商
The Wall Street Journal reported Anthropic expects an operating (EBITDA) profit of $559 million in Q2 2026 on revenue rising to $10.9 billion, but the claim coincides with the startup’s fundraising and lacks public accounting transparency. The article questions Anthropic’s timing and methods, noting its public S-1 disclosures with SpaceX show discounted Colossus compute fees applied in May–June that could temporarily suppress costs and produce a one-off profitable quarter on a non-GAAP basis. Past reporting found higher-than-expected inference costs, and Anthropic plans heavy future spending, so the Q2 profit may not persist. The piece frames the disclosure as PR-friendly financial engineering rather than a durable efficiency breakthrough.
Anthropic says it has reached a multi-year compute deal with an unnamed cloud provider to secure large-scale GPU capacity for its AI model development, addressing training and inference needs for next-generation models. The company framed the agreement as critical to scaling its roadmap and competing with other major labs, highlighting how access to sustained, cost-effective compute can determine which teams can advance frontier models. Anthropic’s announcement matters because long-term, high-volume compute contracts shape competitive dynamics in AI, influence cloud providers’ market positioning, and affect model capabilities, safety testing, and deployment timelines. The deal underscores the strategic importance of infrastructure partnerships in the race to build more capable, safe AI systems.
Google DeepMind and Google Research today unveiled Gemini 3.5 Flash, a new family member designed for agentic workflows and coding that balances frontier intelligence with low latency. The model is available broadly via the Gemini app, Google Search’s AI Mode, Google Antigravity, Gemini API in AI Studio and Android Studio, and enterprise offerings; a higher-capability 3.5 Pro is rolling out next month. Google claims 3.5 Flash matches flagship-level intelligence on benchmarks (Terminal-Bench 2.1, GDPval-AA, MCP Atlas, CharXiv) while delivering up to four times the token throughput and lower cost per task. Integrated with the Antigravity agent platform, it automates multi-step workflows, codebase modernization, web UI/graphics generation, and rapid prototyping, positioning it as a tool for developers and enterprises pursuing large-scale, agent-driven automation.
A new Slack webhook bridge template for Anthropic Claude-managed agents has been added to the claude-cookbooks repo, providing a stateless integration that maps Slack app_mention events into Claude Managed Agent (CMA) sessions and routes session.idle webhooks back to Slack threads via chat.postMessage. The change includes example env variables, a README (CLAUDE.md) with setup and debugging guidance, and files for webhook handling, agent setup, and Slack event wiring. It highlights guidance to use the /claude-api skill as the authoritative API reference and lists optional extensions (GitHub repo mounting, MCP tools, outcomes, multiagent setups, memory stores, custom tools). This matters to developers wanting a production-ready Slack <> Claude agent bridge and accelerates integration and debugging.
Anthropic released Claude Code v2.1.146, renaming the /simplify command to /code-review with an optional effort level and making several fixes and reliability improvements. The update restores AskUserQuestion behavior in Auto mode, fixes Windows PowerShell and Windows Terminal streaming issues, addresses pagination bugs in MCP endpoints, and resolves multiple background session, theme editor, and permission re-prompting problems. It also patches agent SDK streaming exceptions, policy enforcement gaps for third-party and API-key sessions, GNOME paste behavior, and subagent model forwarding. Performance and updater robustness were improved, including retrying downloads on transient network failures and faster diff rendering for large edits. The release targets developer and user experience stability for Claude Code.
Google used its I/O 2026 developer keynote to push an agent-first developer stack: Gemini 3.5 models, Antigravity 2.0 and CLI, and managed Antigravity agents via the Gemini API to simplify agent provisioning and sandboxed execution. Google AI Studio adds native Kotlin, Workspace integrations, and one-click Cloud Run deployment, plus export to Antigravity. Android-focused tooling includes an Android CLI, open-sourced Android skills, Android Bench benchmark with open-weight models like Gemma 4, and a migration agent to convert apps to native Kotlin. For the web, Google proposed WebMCP to let browser-based agents call structured tools and launched Modern Web Guidance to improve performance, accessibility, and security for agent-assisted web development.
At Google I/O 2026, Google unveiled Antigravity 2.0, transforming its coding environment into a full agent development platform that orchestrates specialized subagents and adds an Antigravity CLI, SDK, and managed agents via the Gemini API. Google also announced Gemini 3.5 models, tighter Google AI Studio integrations (including native Kotlin and one-click Cloud Run deployments), and an Android-focused toolchain: an Android CLI, open-sourced Android skills, Android Bench LLM leaderboard, and a Migration agent that converts apps to native Kotlin. On the web side, Google proposed WebMCP, an open standard for browser-executable tools, plus Modern Web Guidance for agent-aware best practices. These moves aim to streamline building, deploying, and hosting agent-driven apps across cloud, Android, and browsers.
A Karpathy-inspired one-file guideline, packaged as a Claude Code plugin and open-source CLAUDE.md, aims to improve LLM coding agents by enforcing four principles: Think Before Coding, Simplicity First, Surgical Changes, and Goal-Driven Execution. The guide addresses common LLM pitfalls—silent assumptions, overcomplication, unintended edits, and vague success criteria—by requiring explicit assumptions, minimal implementations, targeted edits, and test-driven/verifiable goals. It’s available as a Claude Code marketplace plugin (forrestchang/andrej-karpathy-skills) or downloadable per-project CLAUDE.md, and is positioned to make Claude-based coding agents safer, more predictable, and more engineering-friendly. For teams using Claude or similar coding assistants, it offers a practical policy to reduce erroneous or bloated code generation.
Google unveiled Managed Agents in its Gemini API at Google I/O, offering one-call agent deployment that bundles model, harness, sandbox and secure execution into a Google-managed environment. Available in preview via Google AI Studio templates and paired with the Antigravity CLI, the feature aims to eliminate weeks of setup work for execution environments and tool wiring. The move contrasts with rivals—Anthropic embeds orchestration at the model layer, while AWS adds managed harnesses on Bedrock AgentCore—and signals Google’s push for a vertically integrated execution layer. Supporters say it speeds iteration; critics warn it replaces deterministic runtimes with probabilistic services, raising risks around unpredictability and potential data issues.
Ina Fried / Axios : SpaceX S-1: Anthropic is paying SpaceX $1.25B/mo. until May 2029 under their compute deal; Anthropic says it's expanding the deal to include Colossus 2 capacity — Anthropic is paying SpaceX $1.25 billion per month through May 2029 as part of the massive compute deal the companies signed earlier this month.
Google I/O 2026’s opening day delivered a flood of AI and developer-focused product updates, led by multiple Gemini family expansions and a new unified platform. Key announcements included Gemini 3.5 (starting with Gemini 3.5 Flash), Gemini Omni, Gemini Spark, and a broad “Antigravity” platform with its Antigravity CLI replacing the older Gemini CLI. The author, a GDE attendee with front-row access, describes hands-on setup on a Chromebook and notes tighter integration across Google’s stack—Android, Chrome OS Flex, Looker MCP, OAuth, and cross-agent/cross-cloud A2A efforts—while flagging migration pains and strategic consolidation. The changes matter because they push Google toward a single, cohesive developer and enterprise AI platform, shaping tooling and cloud interoperability.
Google announced Antigravity CLI at I/O 2026 and is sunsetting the open-source Gemini CLI, shutting community-run access after June 18, 2026. Gemini CLI, which amassed 100,000+ stars and some 6,000 merged PRs, will be replaced by Antigravity — a Go-built, agent-first, proprietary desktop tool integrated with Google Cloud and designed for multi-agent, enterprise workflows. The author argues this follows a recurring pattern: major tech firms incubate open projects to build community and validate ideas, then pivot to closed, monetized successors that block self-hosting. While Antigravity offers real technical improvements (lower latency, unified backend, security features for enterprises), the change raises concerns about open-source contributors’ work being absorbed into proprietary platforms. This matters for developer trust, governance, and the sustainability of community-led projects.
Google at I/O 2026 unveiled the Gemini 3.5 model series plus a major agent-first push centered on Antigravity 2.0, Antigravity CLI, and managed agents in the Gemini API, aimed at automating complex developer workflows. The company showcased new Google AI Studio integrations (native Kotlin support, Workspace links, one-click Cloud Run deploys, Firebase support) and tools for building Android and web apps, plus security features like sandboxing, credential masking, and hardened Git policies. Google also emphasized tighter cloud and device integration for agents, and announced AI Ultra pricing at $100/month for advanced capabilities. This matters because it stitches advanced LLMs, developer tooling, and managed infrastructure to accelerate app development and production-grade agent deployment across Google’s ecosystem.
百时美施贵宝将部署Anthropic公司的Claude人工智能模型以加速药物研发
@fxtrader: AI领域顶级研究员，OpenAI创始成员及前特斯拉AI总监Andrej Karpathy宣布将入职Anthropic，加入预训练团队，负责Claude大模型的预训练工作。 https://t.co/BjkY4FJj2k
Andrej Karpathy, former OpenAI co-founder and lead researcher, has joined Anthropic, marking a high-profile hire for the AI startup. Karpathy’s move amplifies Anthropic’s engineering and research credibility and signals intensified competition among leading generative AI labs. The hire could influence product positioning, talent recruitment, and market perception even if immediate commercial impacts are unclear; Karpathy’s expertise in large models, system design, and developer-facing products may accelerate Anthropic’s technical roadmap and ecosystem efforts. For rivals and customers, the appointment underscores the ongoing consolidation of top AI talent around a few well-funded labs, which matters for partnerships, model differentiation, and hiring dynamics across the sector.