Small-c context: talking about AI without the jargon getting in the way

Context, tools, capabilities: AI frameworks have claimed the words we use to describe agents. Here's the small-c convention I use to keep them apart.

One thing I've found hard in recent months is talking about general concepts without tripping over words that now mean something specific in AI. A few that keep causing confusion:

  • “We need to review the context this agent operates in.” I mean the overall environment, not its Context window.
  • “Review the tools this agent can use.” I mean everything it can call on, not just its Tools.
  • “We need to define the capabilities this agent will need.” Again, overall functionality, not just PydanticAI Capabilities, even if we're using that framework.

My fix is to use “small-x” for the general concept and keep the capital for the mechanism: “small-c context” is the environment, “capital-C Context” is the context window. Same for “small-s skills” and “capital-S Skills”. It's borrowed from small-c conservative, and once explained it works well in conversation. Not had to use “small-a AI” yet, but I am sure I will at some point.