Chris Schulz·Sep 20From Research Assistant to Constrained AutonomyOn the evidence in this series, an institution can defensibly run agents today for research, filing and document synthesis, hypothesis…
Chris Schulz·Sep 19The Action Surface Is the Attack SurfaceA prompt injection took a trading agent from 7.81% return to 0.89% over a single month. Its Sharpe ratio fell from 5.72 to 0.29. Its trade…
Chris Schulz·Sep 18What the Agentic Trading Evidence Actually ShowsTake ten frontier LLMs, let each run a trading agent on the CSI300 for 548 trading days, and one returns 85.29% against a buy-and-hold…
Chris Schulz·Sep 6Free Claude Code: A Practical Setup Guide, Costs, and Trade-offsLearn how Free Claude Code connects Claude Code, Codex, and Pi to cloud or local models, what the setup involves, and what ‘free’ actually…
Chris Schulz·Aug 23EU Asset Management Regulation in 2026: Five Trends Reshaping the Operating ModelEU asset-management regulation is moving through implementation rather than wholesale redesign. Managers are absorbing concrete…
Chris Schulz·Aug 16EU AI Act Transparency Obligations: A Guide for Providers and DeployersThe EU AI Act does not impose one generic transparency notice on every artificial-intelligence system. It assigns different controls to…
Chris Schulz·Aug 7Amazon Agent Registry in Practice: Discovery, Governance, and Operational LimitsThe Agent Sprawl Problem Is a Discovery Problem First
Chris Schulz·Aug 2Streamlining AI Agent Operations with the AWS Bedrock AgentCore CLIBuilding an AI agent with LangGraph, CrewAI, or LlamaIndex is easy enough on a laptop. Moving that agent to a production AWS environment is…
Chris Schulz·Jul 28The Python Forecasting Stack in 2026: Which Library Fits Which Job?Python has no default forecasting stack. It has mature statistical packages, unified framework APIs, focused high-performance libraries…