Manifesto

AI without context is like a highly capable person whose memory system is impaired.

The motor skills work. Speech works. Reasoning is often still strong. They can write, calculate, analyze, and execute instructions. But every interaction begins almost from zero.

So they ask. Who are you? What are we doing? Why are we doing it? Bit like amnesia, right?

And don't get me wrong, there's nothing wrong with asking questions. But if it's the 10th, the 100th, or the 1000th time?

There's a second layer here too. When you know someone doesn't understand the situation as a whole, you don't trust they'll do the job well. And you're right not to.

In other words: the theoretical ability to execute any task is not the same as being able to execute it well.

The same applies to AI. Even the most fabulous model (Fable included) is non-effective without context it actually understands. And gathering that context is a big job. A single MCP call by an LLM or agent isn't enough in most cases. It'd be like bringing a brilliant, razor-sharp intern into your company, asking them to prepare next quarter's strategy, and allowing them exactly one question before they start.

Not to mention it'd be a nightmare for the intern to even decide which question to ask.

So we decided to spend the next few years on this one problem, solely. How to gather context. How to make AI understand it. How to do it effectively, without it eating your time. And how to make the whole thing seamless.

How to give your intern founder-level understanding of your business. Instantly, the moment you hire them.