The skill right now is curating context for AI. The skill coming is designing knowledge so agents can leverage it. Building personal and organizational data structures that are specifically designed for AI agents is how you proactively prepare for what’s next.

10/ We’re still early. These are the worst AI agents we’ll ever see, and they’re only going to get better. Start thinking about your knowledge infrastructure now. The organizations building it today are the ones agents will actually be able to work with tomorrow.

9/ The real opportunity is thinking about knowledge capture at every level, individual, team, organization, in a way that’s designed to be navigated and used, not just stored.

8/ There are early signals of where this is heading. Tools that use structured files to give agents procedural knowledge they can reference on demand. But that’s one piece of a much larger puzzle.

7/ Which means the question stops being “how do I engineer this interaction” and starts being “is my knowledge even structured in a way an agent can navigate and use?” Most people aren’t thinking about this yet. Most organizations aren’t either.

6/ Context engineering is still a human-driven act. You decide what to bring, you package it, you get an output. It works. But as agents become more embedded in how work gets done, that model starts to break down. Agents don’t wait for you to hand them context. They go find it.

5/ Humans and I need different context. Curated context for humans and curated context for AI agents are going to look very different. Building for one doesn’t mean you’ve built for the other.

4/ Now there’s a next layer forming, and I think it’s the most important one yet. Context architecture. Not just how you package context for a single interaction, but how you design and structure your knowledge so it’s modular, reusable, and actually usable when it’s needed.

3/ The sharper practitioners started shifting toward context engineering. Less about the perfect prompt and more about what you bring into the interaction. It’s a small shift in mindset that changes everything about what you can actually do with AI.

2/ From there, the focus shifted to prompt engineering. Crafting better inputs, building prompt libraries, sharing techniques to get more consistent outputs. A real skill, and still useful. But a lot of people stopped here.

1/ When ChatGPT launched, the first wave was just trying to understand what we were looking at. What is an LLM? What can it actually do? That foundational understanding is still where a lot of people and organizations are, and it’s a reasonable place to start.

The organizations that win with AI agents aren’t going to be the ones with the best prompts. They’re going to be the ones with the best knowledge infrastructure. Here’s what I mean. 🧵

Most AI users only use it as a Google replacement. Instead of searching for answers, use it as a thinking partner. Use voice-to-text and just think aloud. Explain a topic like you would to a friend or colleague. You’ll learn a lot about new amazing ways of working with AI.

Saw a job posting that required 5+ years of experience using ChatGPT. ChatGPT was released to the public December 2022. Job seekers - good luck on that.

Found my first cd I burned 20+ years ago. Track 1: The Anthem by Good Charlotte. Eccentric mix of burnt cds just hits different than the algorithmically curated music streaming we have today. Maybe it’s just nostalgia speaking, but there’s something special about it.

The future of work won’t be AI vs humans. It’ll be humans who know how to work with AI vs. those who don’t.

Reduce the time from idea to execution

Not doing the thing makes it harder to do the thing when you finally decide to do the thing So just do the thing you know you need to do that you have not done yet

Find people that have gotten to where you want to go. Observe them. Do what they do. Ask really good questions. Do the hard work each day. Over time, small daily habits will compound into transformational results.

You can do anything But you can’t do everything Strategy is prioritizing what to do