Core thesis
Intelligence is becoming abundant. Context, judgment and trust are not. Our thesis, and the doubts we keep beside it.
From the desk2 min read
Here is the short version. As intelligence becomes cheap and everywhere, the scarce things are the ones a model cannot supply on its own: knowing what matters to you, judging what to do about it, and being trusted to decide. We build for those scarce things.
Side A · The thesis
Context beats raw capability.A general model knows a little about everything. A good tool knows a lot about you: your questions, your sources, your standards. That context is what turns an answer into a useful one.
Judgment stays with people.The best results come from tools that do the reading and leave the deciding. Every instrument we build shows its work, so the person using it can check it.
Tools should be kept, not rented.What you learn should stay with you: the notes, the highlights, the conclusions. We design for ownership, so the value compounds for the person and not only for the platform.
Operating businesses make better tools.Cityscope and Align Works run real operations every day. They are the first place our instruments are tested, and they keep us honest about what works.
Accessible beats exclusive.Knowledge work is not only analysts and engineers. It is students, creators, researchers and anyone learning in public. The tools should meet them where they are.
Intelligence is becoming abundant. Judgment is not.
Side B · The doubt
House ruleEvery Analog Research piece carries a thesis side and a doubt side. If a view cannot name its weak spots, it is not finished.Context could become a commodity too.If the largest platforms learn enough about everyone, a small team's edge narrows. Our answer is to stay close to specific people and specific work.
People may not want the final call.Convenience is powerful, and some people will prefer an answer to a judgment. We think the ones who want to understand will matter most, but we hold this loosely.
Holding companies can lose focus.Running operating businesses and building instruments at the same time is hard. We keep the portfolio small on purpose.
What would change our mind
If people consistently trust tools that hide their sources over tools that show them, or if owning their notes and knowledge stops mattering to the people we build for. We will report on both, here, as we learn.
End of transmission