Next version
Run: AI reading the context you built
Start collects, organises and calibrates. Run interprets. This page shows what comes next, and says clearly what is already decided and what is still vision. Nothing here exists in today's version.
Why Start comes first
An AI with no data of your own is noise. An AI connected to a badly calibrated context fires blind. So Mynd starts without AI: the person first builds their context, with their judgement, and Start already records the indicators and how they vary over time. When Run arrives, half the collecting is done.
What is different in Run
Most personal AIs look at content: documents, conversations, messages. Run also looks at people and relationships, from the ratings and the changes Start already keeps.
- Read the numbers in Self together and say what changed from one week to the next.
- Suggest tags and Context phrases for records not yet filed.
- Warn about stalled subjects and reminders always put off.
- Rebuild the story of a decision from the activity log.
- Write the summary at the top of an automatic report.
- Automatic sends without today's limit: Start has 14 search sends and 14 card sends; in Run, each person will have as many as they need.
Principles already decided
- It analyses the data, never the person. Run works on the Digital Self, the representation the person built, not on the human being.
- The person decides what leaves. Mynd represents the person; it never hands anything to a third party who asks for it.
- No inflating, neither to please nor to seem tough. A document generated by Run says only what the data supports.
- Off by default. The AI connection only exists if the person switches it on, with a clear warning that the chosen texts are sent to Anthropic. The database stays on the phone.
- One AI. Run comes with Claude, by Anthropic, built into the app, without mixing providers.
- Time weighs in. Analyses look at several periods, from the last week to the last years, each with its own weight.
Still being designed
- Portraits of the Digital Self by domain (work, business, relationships), which the person can use with third parties, by their choice.
- A structure of its own for the Digital Self, separate from the structure of the records.
- Validation of the analysis criteria with psychology professionals, before any conclusion about people.
After Run
A further phase is designed, with several AIs coordinated around the same context, each with its own role and a supervisor above them, and the person always as the final supervisor. Without a common context, coordinating several AIs would be chaos; Mynd is what gives them coherence. It is a vision on its way, conditional on getting there.
Mynd Start works fully without any of this. Whoever buys today buys Start.
