Amazon enforces a 90-day safety guardrail on AI-generated code for consumer-facing systems following costly outages, requiring stricter documentation and dual approvals, as experts warn that AI code often appears production-ready before being properly tested amid an industry-wide AI coding rush.
Claude's Excel and PowerPoint add-ins now share full context across all open files in one continuous conversation, offer one-click Skills for tasks like LBO models and pitch deck reviews, and are deployable across Amazon Bedrock, Google Vertex AI, and Microsoft Foundry.
Agile movement pioneers reunite in Utah for a landmark retreat examining how AI is fundamentally reshaping software engineering, raising urgent questions about human value, security risks, and the future of coding as AI agents take over code generation.
Cursor expands its Marketplace with 30+ new plugins from major partners including Atlassian, GitLab, Datadog, and Hugging Face, giving developers powerful new ways to manage infrastructure, boost productivity, and automate workflows using natural language directly within their coding environment.
Cloudflare launches a powerful new /crawl endpoint in open beta, enabling developers to crawl entire websites with a single API call and retrieve content in HTML, Markdown, or structured JSON formats, with advanced controls like depth limits, URL filtering, and incremental crawling available on both free and paid plans.
AI coding tools are dramatically closing the performance gap in software engineering, boosting struggling teams' delivery speed by nearly 50%, yet top performers still dominate — shipping software in under 22.5 days compared to over 62 days for bottom-tier teams, as experts predict AI will fundamentally reshape team structures worldwide.
GitHub launches its Copilot SDK, giving developers direct access to the same production-tested agentic AI engine powering GitHub Copilot, enabling autonomous multi-step task execution, tool invocation, and file modification inside their own applications without building custom orchestration stacks.
A viral Forbes claim that Anthropic loses $5,000 per Claude Code Max user is misleading, as actual compute costs are estimated at just $500 — and on average, Anthropic is likely profitable on most Claude Code subscribers, with typical developers costing only $18 per month to serve.
Apple's M5 Max MacBook Pro, with 128GB unified memory, is enabling developers to run massive 70-billion-parameter AI models locally for free as cloud AI inference costs spiral out of control, with some startups reporting costs tripling in just three months.
Apple's MLX framework now supports the M5 chip's Neural Accelerators, delivering a massive 4x speedup in AI response times and up to 27% faster token generation, powered by the M5's 153GB/s memory bandwidth, enabling MacBook Pro users to run powerful large language models locally with ease.