Tech • IA • Crypto
Anthropic Claude Routines, AWS Bedrock, Devin reshape coding
Briefing
Claude Routines enable proactive agents
Anthropic introduced routines in Claude Code, turning it from a reactive tool into an always-on agent. Developers can define prompts, tools, repos, and triggers, while infrastructure is fully managed. Routines run on schedules or events like GitHub activity and webhooks. The shift removes cron jobs and custom backends, pushing AI toward autonomous execution.
AWS doubles down on Anthropic
Amazon Web Services deepened its partnership with Anthropic, pairing Claude with Amazon Bedrock and new deployment paths. The initiative includes Project Rainier and third-gen Trainium chips to cut cost and boost performance. Anthropic has committed over $100 billion in cloud usage, signaling long-term alignment. Enterprise features emphasize zero operator access and strict data isolation.
Microsoft Foundry integrates Claude models
Microsoft Foundry is positioning itself as a central hub for agentic AI development with Claude Sonnet 4.6 and Opus 4.7. The platform connects with VS Code, GitHub, and over 1,400 connectors including SAP and ServiceNow. It focuses on production readiness with governance, observability, and orchestration. The move highlights growing enterprise demand for multi-step reasoning systems.
Devin signals autonomous coding shift
Systems like Devin represent a leap from assistants to fully autonomous AI software engineers. These agents can plan, write, and execute complex codebases with minimal supervision. The ambition is to make development up to 10× faster across industries. Adoption is spreading in sectors like finance, healthcare, and government.
Spotify and Binti show real impact
Spotify uses AI agents to migrate repositories, generating over 1,000 pull requests per month and cutting timelines by 90%. In social services, Binti reduced foster care licensing delays by 20 days using automation. These deployments highlight tangible productivity gains beyond prototypes. AI agents are increasingly embedded in operational workflows.
Claude Code gains autonomy features
Anthropic upgraded Claude Code with more independent execution and remote workflows. The system now evaluates action safety, proceeding without user approval for low-risk tasks. A new remote control feature allows sessions across devices with persistent state. Interface improvements include a flicker-free terminal and better session organization.
Base44 scales with AI workflows
Base44 grew from a solo founder to 80 engineers while maintaining rapid development via AI automation. The company replaced documentation with AI-generated onboarding and system analysis. Code reviews and testing were automated using patterns from past decisions. After launching in 2024, it reached profitability by April 2025 and integrated with Wix.
Test-time compute reshapes performance tradeoffs
Increasing test-time compute—allowing models to use more tokens—boosts reasoning and accuracy. Gains are significant across coding and scientific tasks but show diminishing returns at high levels. The tradeoff between cost, latency, and intelligence is becoming central to deployment decisions. This dynamic mirrors earlier scaling laws but shifts focus to inference-time optimization.
Videos covered
The Problem Solvers: Scott Wu at Cognition
- •Rise of AI Software Engineers
- •Mission to Accelerate Development
- •Foundations in Competitive Programming
Code with Claude London 2026: Opening Keynote
- •From hobbyist coding to autonomous systems
- •AI agents delivering real-world impact
- •Rapid acceleration in model capabilities
Build a proactive agent workflow with Claude Code
- •Shift from reactive tools to proactive agents
- •Infrastructure challenges addressed
- •Introduction of “routines” feature
Build AI agents using Claude in Microsoft Foundry
- •Shift to Agentic AI Systems
- •Microsoft Foundry as a Unified Platform
- •Claude Models for Advanced Reasoning
AI with Claude on AWS: From code to orchestration
- •Deepening AWS–Anthropic partnership
- •Project Rainier and custom AI chips
- •Three deployment paths for Claude on AWS
What's new in Claude Code
- •Shift Toward Autonomy
- •Remote Control Across Devices
- •Flicker-Free Terminal Interface
From one person to 80: Scaling a hypergrowth engineering org with Claude Code
- •Rapid early growth and profitability
- •Scaling from 1 to 15 engineers
- •AI-driven onboarding replaces documentation
The thinking lever
- •Test-Time Compute Boosts Performance
- •Token Usage Directly Impacts Output Quality
- •Diminishing Returns at Maximum Effort