Devlery

Blog

Notes and analysis on AI development.

AWS MCP Packs 15,000 APIs Into a New Boundary for Cloud Agents

AWS MCP Packs 15,000 APIs Into a New Boundary for Cloud Agents

AWS Agent Toolkit and the AWS MCP Server GA show how coding agents can reach cloud accounts through IAM, CloudWatch, and CloudTrail.

Docusign MCP Beta Turns Agreements Into Agent Tools

Docusign MCP Beta Turns Agreements Into Agent Tools

Docusign Iris Agents and its MCP beta show how agreement data can become a callable work surface for Claude, Gemini, and ChatGPT.

Small browser agents cross Operator's 58% line

Small browser agents cross Operator's 58% line

Microsoft Fara1.5 shows why small computer-use models, synthetic web environments, verifiers, and approval policies are becoming the browser-agent bottleneck.

The 15x token bill and the return of the AI-native cloud

The 15x token bill and the return of the AI-native cloud

DigitalOcean AI-Native Cloud shows why agent costs are shifting from GPU rental to inference routing, data, state, and operations.

The 24-Hour Agent Permission Problem in Front of 900M Gemini Users

The 24-Hour Agent Permission Problem in Front of 900M Gemini Users

Google Gemini Spark brings background agents, MCP connections, and approval boundaries into a mass-market consumer AI surface.

99.82% Cache Hits, the New Variable in Coding Agent Costs

99.82% Cache Hits, the New Variable in Coding Agent Costs

The Reasonix debate shows that coding agent costs depend not only on model pricing, but on harness design that keeps prefix cache intact.

20 missed attacks, SLEIGHT-Bench warns agent security teams

20 missed attacks, SLEIGHT-Bench warns agent security teams

SLEIGHT-Bench uses 40 synthetic attacks to show how easily LLM monitors can miss risky behavior by coding agents.

The coding agent market gets a new enterprise scorecard

The coding agent market gets a new enterprise scorecard

Gartner and OpenAI show how AI coding agents are moving from model benchmarks toward governance, sandboxing, auditability, and cost control.

Copilot Remote Control Sets New Rules for Agent Costs

Copilot Remote Control Sets New Rules for Agent Costs

GitHub Copilot CLI remote control and automatic model routing turn coding agents from IDE helpers into long-running sessions that teams must operate.

x402 hits $24.24M, and agent wallets now need accountability

x402 hits $24.24M, and agent wallets now need accountability

Circle Agent Stack opens a path for AI agents to pay for APIs and services with USDC, while exposing sharper responsibility and security boundaries.

15B tokens per minute, the real cost of OpenAI’s superapp strategy

15B tokens per minute, the real cost of OpenAI’s superapp strategy

OpenAI’s $122B raise is not just financing. It ties ChatGPT, Codex, API usage, enterprise adoption, and compute into one agent-first flywheel.

Two H100s Are Enough, Command A+ Targets Private Agents

Two H100s Are Enough, Command A+ Targets Private Agents

Cohere Command A+ lowers the bar for private AI agents with Apache 2.0 open weights and a two-H100 deployment target.