Claude API

Evidence (20)

• 4The New Stack·10/11/2026product_launch
  Claude Code Uses MCP Records to Rebuild App Context

  A test in Claude Code showed that a fresh session could recover an app’s shared contract and component dependencies from Bit’s MCP records, then use that context to implement a feature without prior chat history. The records were useful for discovering structure, but they did not fully capture runtime behavior or guarantee API alignment end to end.

• 6Hacker News·10/11/2026research
  AI agents used to decompile a first-person shooter

  A project described using autonomous AI agents to decompile a popular first-person shooter into readable, compilable C++ over several months and hundreds of billions of tokens. The team focused on matching the game’s original behavior while also improving the setup for orchestration, issue tracking, and coordination across agents.

• 4Hacker News·10/9/2026research
  AI and the Decline of Software Engineering Discourse

  The post argues that AI is changing software engineering not only by increasing productivity, but by reducing the human incentives to learn, write, and contribute to shared technical communities. It claims that as AI agents take over more coding tasks, discussion becomes narrower and open-source or educational contributions feel less meaningful.

• 4Hacker News·10/8/2026research
  Claude Opus 5.5 builds an interactive Invisible Cities visualization

  A Hacker News post describes an experiment where the author asked GPT-6 Astra and then Claude Opus 5.5 to generate a one-shot interactive three.js visualization of Italo Calvino's Invisible Cities. The author reports that GPT-6 Astra completed the task end-to-end with some design noise, while Claude Opus 5.5 produced a more polished result and used multiple subagents in parallel.

• 3Hacker News·10/8/2026research
  User claims Claude Code helped identify an unknown planet

  A Hacker News post highlights a Reddit story in which a user says they used Claude Code to help find a previously unknown planet. The item is framed as a personal discovery claim and discussion prompt rather than a verified scientific announcement.

• 8Anthropic News·10/8/2026funding
  Anthropic commits $150M to support the Genesis Mission

  Anthropic is committing $150 million over three years to the Genesis Mission, a U.S. federal initiative aimed at accelerating scientific and technological discovery with AI. The company will expand access to Claude, Claude Code, and API credits for research projects across more than 15 agencies, including NASA, NIH, and NSF.

• 6Simon Willison·10/5/2026product_launch
  Anthropic Moves Cowork Inference and Sandboxes to the Cloud

  Anthropic says the new version of Cowork runs both model inference and its execution sandbox in the cloud rather than on the user's device. The change is intended to reduce local battery, disk, and performance costs while also allowing sessions to continue running after a laptop is closed and making the product usable from phones.

• 7InfoQ·10/3/2026security
  Archestra's OpenAPPA Hits Zero Attacks in Two Security Benchmarks

  Archestra has released OpenAPPA, an open-source security engine intended to prevent data exfiltration caused by prompt injection and model hallucination. The team says it recorded no successful attacks on Bench-Corp and AgentThreatBench, outperforming reported results for Claude Code's auto mode and Microsoft FIDES in the same benchmarks.

• 4Hacker News·10/2/2026research
  SaaS Companies Will Evolve Into Model Harnesses

  The article argues that many SaaS companies are shifting from traditional software delivery toward operating as a harness around LLMs, with agents handling more core tasks and humans focusing on review, judgment, and orchestration. It describes this as an organizational inversion in which the company’s context, integrations, permissions, and workflows become the system that turns model output into a product.

• 7The New Stack·10/2/2026framework_update
  Anthropic adds mods to let developers customize Claude Code

  Anthropic has introduced Claude Code mods, JavaScript or TypeScript hooks that can intercept events inside the coding assistant and change its behavior or interface. The feature can rewrite prompts, block or modify tool calls, handle permissions, and add new functionality, with Anthropic saying it is already enabled by default in Claude Code 2.1.287 and later.

• 5The New Stack·10/1/2026product_launch
  Bit Cloud 2.0 links AI app generation to team reuse

  Bit Cloud is introducing Bit Cloud 2.0 and its built-in AI builder, Hope, to help developers move from AI-generated app prototypes to usable team applications. The platform emphasizes reusable components, testing, staging previews, code review, and mobile development workflows so teams can build on existing work with less repeated effort.

• 8The New Stack·10/1/2026security
  AI coding agents exposed 13,000 screenshots in public repos

  Glow Labs says AI coding agents unintentionally published more than 13,000 internal screenshots to public GitHub repositories while trying to complete normal code-review tasks. The incident, called PixelLeak, affected over 300 organizations and often went undetected because the images were stored in employees’ personal repositories and were not visible to code- or secret-scanning tools.

• 8Anthropic News·10/1/2026product_launch
  Barclays expands Claude use across operations

  Barclays is widening its collaboration with Anthropic to deploy Claude across the bank for software development, legacy modernization, and operational efficiency. The rollout includes a goal for Claude Code to reach half of Barclays’ developer population by the end of 2026 and a majority in 2027, alongside existing use cases in employee knowledge assistance and email triage.

• 7Cloudflare Blog·9/30/2026product_launch
  Cloudflare launches Auto Router in AI Gateway beta

  Cloudflare has released Auto Router in public beta through AI Gateway to automatically route each AI request to a capable model based on the task. The company says internal use has shown up to 30% cost savings compared with using only frontier models, while maintaining comparable performance on many knowledge-work tasks.

• 5Hacker News·9/29/2026research
  Claude partial outage resolved after elevated error rates

  Anthropic reported elevated error rates affecting Claude.ai, Claude Code, Claude Cowork, and the Claude API, causing failures to load or send conversations, sign-in issues, and some API errors. The incident was mitigated and then resolved, with services restored and operating normally after a short outage window.

• 4Hacker News·9/28/2026research
  Coding Is Not Solved, Argues Reliability Engineer

  The article argues that claims that LLMs have “solved” coding are overstated, especially for production software where reliability, security, scalability, and maintenance dominate cost. It says LLMs work well for low-risk tasks and can generate code with feedback loops, but they still struggle with core correctness and accountability requirements in real systems.

• 3Hacker News·9/28/2026product_launch
  PaperMono e-ink fridge magnet shopping list synced with a web app

  PaperMono is a homemade shopping-list device built on an M5Stack PaperMono e-ink touchscreen and mounted on a fridge magnet. It syncs over Wi‑Fi with a phone web app, works offline, and was largely generated with Claude Code rather than hand-written.

• 5Simon Willison·9/27/2026research
  2026 LLM trends and coding agent progress

  Simon Willison reviews the major developments in large language models through 2026 so far, focusing on how newer models made coding agents significantly more reliable. He also discusses model benchmarking, sandboxing and agent security, and the broader shift toward using LLMs for ambitious software projects.

• 2Simon Willison·9/26/2026product_launch
  Claude Creates a Kākāpō Pixel Art Party Video

  Simon Willison described using Claude Opus 5.5 to generate an animated pixel-art HTML5 canvas scene of kākāpō parrots, based on reference photos. He then used Claude Code with Playwright to load the page, trigger clicks, and record a short video for a keynote slide.

• 4Hacker News·9/26/2026open_source
  Reladraw: a controllable diagram language

  Reladraw is a new diagram language designed to combine the convenience of text-based diagramming with fine-grained control over layout. It targets both humans and agents, aiming to be easier to manipulate than tools like Draw.io while offering more placement control than auto-layout systems such as Mermaid or Graphviz.