Session two of the reading series — continuing the guided tour through agentic AI, online in the evening.
What we covered
With the foundations set, session two went deeper into how agents are actually put together — the patterns that separate a demo from a system.
- Tools and function calling — how agents reach outside the model: schemas, validation, and failure modes.
- Planning vs. reacting — when to decompose up front and when to let the loop figure it out.
- Memory that matters — what to keep, what to summarize, and what to throw away between steps.
What we built
Shared notes and a growing pattern catalogue — 35 learners comparing implementations and war stories from their own experiments between sessions.
Takeaways
- Most agent failures are tool-design failures, not model failures.
- Planning is a spectrum: the right point on it depends on the task, not the fashion.
- Evening online sessions hit different — 35 people showed up mid-week for this one.
Join the next one
We run sessions like this regularly — free, welcoming, and open to everyone. RSVP on Meetup and bring your questions.
