Context, tokens, and working well with AI coding tools
2026-06-18
You use chatbots. You write prompts. You judge whether the answer is any good.
That intuition transfers directly. This session adds three things:
So much of our day is file-shaped:
A chatbot reasons about a snippet you paste. An agent reasons about the files as they actually are.
“Open this repo, find the script that reads the TSV, explain how the columns are normalized, then add a validation check before write-out.”
A capable but junior colleague — not an oracle.
Your job shifts:
You can often swap the model and keep the workflow. Model = intelligence. Framework = the workflow that puts it to work.
Reading is cheap; writing is expensive. The size of the answer matters, not just the size of the prompt.
MCP (Model Context Protocol) — a standard way to connect an agent to external tools and data.
Start small and concrete — not “build a web app.”
Agentic tools are most useful when the work depends on the state of real files, real commands, and real project context.
Token/price/context figures are mid-2026 and illustrative; exact numbers move fast.