Agents are now teammates, users, and builders. SennaMind covers the new engineering practices that make them work: agent loops, backpressure, orchestration, memory, and Agent Experience (AX) — hands-on, with honest failures included.
Software engineering is being rebuilt around agents. These are the disciplines that matter now.
Ralph loops and stateless iteration — letting agents run until the job is done, safely.
Gates, tests, and verification harnesses — the brakes that make autonomous speed survivable.
One goal, many agents — decomposition, parallelism, and keeping humans in the loop.
Teammates that remember — provenance, context engineering, and durable learning.
Agents are users now. Desire paths, llms.txt, MCP — building products agents can actually use.
Each playlist is a learning path — concepts first, then hands-on builds with real agent transcripts.
Working with agents as a team — loops, backpressure, orchestration, and memory in practice.
Your newest user isn't human. Desire paths, and why building for agents is the next UX.
One agent-hostile product, fixed pillar by pillar — from 1/5 to 5/5 on the agent scoreboard.
Build real software with open-source models at $0/month — honest benchmarks, full builds.
I build with agents every day — and SennaMind is where I share what actually works.
The thesis behind everything here: the bottleneck isn't model capability anymore — it's the harness around it. Loops need brakes, agents need context, and products need to treat agents as first-class users.
Reach me at cmm@sennamind.com .