Don't Invest in AI. Invest in Skills.

This is not a text about human skills and how the AI hype ignores them. Skills rarely play a role in hypes—they take time to build, and the organisations that survive the hype are the ones that actually build skill.

But no, it's not about organisational skills either.

I'm talking about a concept in Claude called skills.

Context is Everything

If you've been playing around with prompt-based interfaces to AI models, you know: the more context you give, the better the answer gets. The better you explain what to do, the better the result will be.

Skills are that contextual application of how to approach a certain problem.

A skill is a text document, written in plain language:

name: explain-code

description: Explains code with visual diagrams and analogies.
Use when explaining how code works, teaching about a codebase,
or when the user asks "how does this work?"

When explaining code, always include:

1. Start with an analogy: Compare the code to something from everyday life
2. Draw a diagram: Use ASCII art to show the flow, structure, or relationship
3. Walk through the code: Explain step-by-step what happens
4. Highlight a gotcha: What's a common mistake or misconception

Keep explanations conversational. For complex concepts, use multiple analogies.

(example from the Claude Code website)

With this type of programming, you ensure the AI performs its task the same way every time.

When you create a skill, you program the AI.

The only difference to artificial programming languages? You use your natural language. This removes an important barrier from "programming things"—you're not forced to translate your thoughts into an artificial language. You can speak the language you speak anyway.

Great news—because now literally every person who can speak can become a programmer.

The Real Work

The main job is to translate a task and an approach into programmatic instructions. That sounds intentionally difficult, because it is.

Learning to frame a problem or task in a way that allows for programmatic execution is something you need to learn. It's taught in programming classes—there's no way to program without doing so. But it's also taught in literature and philosophy.

The mental capability of understanding a real-world situation (in legal science we say in German "Lebenssachverhalt") and describing how, from a certain perspective, this situation should be addressed—that's the prerequisite of programming AI.

Where the Value Lives

Now to come back to the flashy title: The value is not in the AI model.

AI models will evolve, become better, benefit from general technological advancement. If you're not Anthropic or Mistral, you don't need to invest in an AI model.

But the thing worth investing in? Skills.

Documenting how your organisation wants to handle a certain real-world situation—as you would explain it to a new joiner of the organisation. This translates into skills which persist across changing AI models. And that makes them a very sustainable investment.

That's why you should not invest in AI, but in skills.

What's your take? Are you documenting how your organisation approaches problems, or are you still chasing the next AI model?

#AI #Skills #Innovation #Programming #KnowledgeManagement