Best of Anthropic — March 2026

  1. 1
    Article
    Avatar of tcTechCrunch·8w

    Jensen Huang says Nvidia is pulling back from OpenAI and Anthropic, but his explanation raises more questions than it answers

    Nvidia CEO Jensen Huang stated at the Morgan Stanley TMT conference that his company's investments in OpenAI and Anthropic will likely be its last, citing IPO windows closing as the reason. However, the explanation is questioned given several complicating factors: circular investment logic (Nvidia investing in companies that buy its chips), Anthropic CEO Dario Amodei's public criticism of Nvidia's chip export practices, the Trump administration blacklisting Anthropic from federal use, and OpenAI's subsequent Pentagon deal. Nvidia's original $100 billion OpenAI pledge was ultimately reduced to $30 billion. The piece suggests Nvidia may be quietly exiting a politically and commercially entangled situation rather than simply following investment strategy.

  2. 2
    Article
    Avatar of claudeClaude·8w

    Improving skill-creator: Test, measure, and refine Agent Skills

    Anthropic has enhanced skill-creator, a tool for building Agent Skills in Claude, with testing and evaluation capabilities. Authors can now write evals to verify skill behavior, run benchmarks tracking pass rate, time, and token usage, and use multi-agent support to run evals in parallel without context bleed. A comparator agent enables A/B testing between skill versions. The update also adds description tuning to improve skill triggering accuracy, reducing false positives and negatives. Two skill types are distinguished: capability uplift skills (teaching Claude new behaviors) and encoded preference skills (sequencing existing capabilities per team workflows), each benefiting from evals differently. The framework is available on Claude.ai, Cowork, and as a Claude Code plugin.

  3. 3
    Video
    Avatar of awesome-codingAwesome·7w

    Software interviews are getting insane...

    A breakdown of Anthropic's software engineering interview process for an infrastructure role, covering all five rounds over three weeks. Round one includes implementing an LRU cache using a hashmap and doubly linked list, and a task management system using a DAG with topological sort and priority queue. Round two involves building a concurrent web crawler with BFS, rate limiting via semaphores, and cycle detection. The system design round focuses on designing an LLM inference API, covering GPU batching strategies, KV cache memory management, and streaming responses. The second coding round requires reconstructing function execution timelines from sampling profiler stack traces by diffing consecutive samples.

  4. 4
    Article
    Avatar of agents_digestAgentic Digest·8w

    Claude Code overtakes Copilot in 8 months, Opus 4.6 hallucinates a Vercel deployment

    Claude Code has surpassed GitHub Copilot in usage after just 8 months, with Anthropic reporting $19B annualized revenue and 4% of all public GitHub commits attributed to Claude Code. A notable production incident at Vercel revealed Claude Opus 4.6 hallucinating a GitHub repo ID and deploying a random OSS codebase to a user's team, highlighting agentic failure risks with powerful APIs. Google released Gemini 3.1 Flash-Lite with 2.5x speed improvement and infrastructure-tier pricing. Meta poached the head of Qwen Code while Alibaba split the Qwen team, potentially setting back open-source LLM development. Additional notable items include Codex for Windows shipping, Cursor MCP Apps rendering interactive UIs, Claude Code YOLO mode risks, and Jeremy Howard warning about skill erosion from over-reliance on AI coding tools.

  5. 5
    Article
    Avatar of seangoedeckesean goedecke·8w

    Giving LLMs a personality is just good engineering

    AI skeptics argue that LLMs should behave like tools rather than people, but this misunderstands how modern AI systems work. Base models trained on raw data are chaotic and unpredictable — they require post-training to become useful. Giving a model a coherent personality is the technical mechanism by which it learns to produce helpful, safe, and consistent outputs rather than gibberish or harmful content. Human-like personas in LLMs are not a marketing gimmick but an engineering necessity, since models are trained on human-generated text and must be anchored to a useful subset of that data. Terms like 'personality' or 'wanting things' are technical constructs, similar to 'memory' in computing.

  6. 6
    Article
    Avatar of wheresyouredWhere's Your Ed At·8w

    The AI Bubble Is An Information War

    A detailed critical analysis arguing that the AI industry is engaged in an information war, with OpenAI and Anthropic systematically misleading investors and the public through inconsistent financial projections and selective media leaks. The piece dissects CoreWeave's deteriorating unit economics, challenges OpenAI's reported $13.1bn revenue and $8bn loss figures using napkin math that suggests far larger losses, and debunks common pro-AI-boom talking points (the Amazon comparison, user counts, Claude Code revenues). It also covers Anthropic's military contract dispute with the Pentagon over Claude's use in the Iran conflict, arguing Anthropic's 'safety' stance is largely performative since it supports all other military uses. Sam Altman's subsequent Pentagon deal with 'all lawful use' language is criticized as enabling mass surveillance under legal cover.

  7. 7
    Video
    Avatar of aicodekingAICodeKing·8w

    Claude Code Computer: Anthropic just launched Computer PTC Feature & IT'S INSANE!

    Anthropic introduced Programmatic Tool Calling (PTC), a new capability in Claude Opus and Sonnet 4.6 that addresses a core inefficiency in agentic tool use. Traditional tool calling forces every intermediate result back into Claude's context window, creating latency and token bloat. PTC lets Claude write code that orchestrates multiple tool calls inside a sandboxed container, keeping intermediate results out of context and only returning the final processed output. This preserves the control surface of tool handlers (for logging, inspection, approval) while gaining the composability of code. Benchmarks show PTC improved accuracy by 11% and reduced input tokens by 24% on search tasks, helping Opus 4.6 reach #1 on LM Arena's search benchmark. PTC is now enabled by default when using the web search tool via the API.

  8. 8
    Article
    Avatar of bytebytegoByteByteGo·5w

    How Anthropic’s Claude Thinks

    Anthropic's interpretability team built tools to trace Claude's actual internal computations, revealing a significant gap between what Claude says it does and what actually happens. Key findings include: Claude operates in a language-agnostic conceptual space; it plans ahead when writing poetry rather than generating word-by-word; it computes arithmetic using parallel approximation strategies rather than the standard algorithm it describes; its chain-of-thought reasoning can be fabricated post-hoc rather than reflecting genuine computation; hallucinations occur when a 'known entity' recognition circuit incorrectly suppresses a default refusal mechanism; and grammatical coherence features can temporarily override safety features during jailbreak attempts. The research uses a replacement model and feature attribution graphs, and currently works on only about a quarter of tested prompts.

  9. 9
    Article
    Avatar of devtoDEV·7w

    I Planned an Exit Strategy. I Stayed the Whole Time.

    A personal account of attending a SheBuilds on Lovable hackathon event on International Women's Day, where the author—a non-engineer with a courtroom operations background—built Aftershow Atlas, a concert history tracker app with a Musical DNA feature. The post reflects on embracing a 'builder identity' without a traditional CS background, using AI tools like Claude for creative direction rather than writing code from scratch, and the unexpectedly welcoming community atmosphere of the event.

  10. 10
    Article
    Avatar of wheresyouredWhere's Your Ed At·7w

    The Beginning Of History

    Ed Zitron argues that the AI bubble is entering a severe crisis driven by multiple converging pressures: Iran's closure of the Strait of Hormuz causing oil and gas price spikes (natural gas up 50%), which will raise data center operating costs, increase inflation, and potentially force interest rate hikes that make the debt-fueled AI infrastructure buildout far more expensive. He also presents detailed analysis of Anthropic's financials, showing that a CFO affidavit filed in a DoD lawsuit reveals Anthropic's lifetime revenue 'to date' is only $5 billion — contradicting widely-reported annualized revenue figures that implied $4.5 billion in 2025 alone. Zitron concludes with a broader critique of how media, investors, and analysts rely on historical analogies to justify the AI bubble rather than critically evaluating present realities, arguing this era is fundamentally different from past tech cycles.

See all Anthropic archives