Computer as instrument, not lifestyle product

There are two lineages of professional software.

One descends from the consumer web. It optimises for delight, onboarding speed, viral loops, and the satisfaction of crossing items off a checklist. Notion, Linear, Airtable. They are good at what they do. They are not what we are building.

The other descends from instruments. Bloomberg. IBM mainframes. The trading turret. Reuters. Modelling tools that survived three decades because they did not chase the consumer’s attention — they served the operator’s hands. They are taught, not discovered. They are configured, not onboarded. They reward fluency, not novelty.

Yig sits in the second lineage.

What this means in practice

Someone finishing the monthly update at 11pm on the last day does not need a confetti animation when a change lands. They need the instrument to be there, predictable, fast, and quiet. They need to trust that what they did last period still works this period. They need the audit trail to write itself.

Lifestyle products optimise for the moment. Instruments optimise for the decade.

That decision cascades into every choice we make:

• No dashboard. Dashboards are consumer ergonomics borrowed for professional problems. The person who signs off does not want to log in to one more place. The work stays in the Excel workbooks the team already uses, and Excel computes every figure.
• No theatrical AI. No streaming animations of the agent “thinking”. No cute personality. The agent does its work, presents the result, gets out of the way.
• No retraining the operator. An operator with twenty years of muscle memory in their workbook is a feature of the team, not a problem to solve. Yigcore works with the workbook; it does not replace it.
• Explicit over implicit. Every output is reviewable. Every change is logged. Nothing happens “automatically in the background” that the operator did not see.

The bet

We are betting that the people who own these workbooks prefer instruments to toys, even when the toys are very impressive. We are betting that what looks slow now — disciplined integration into existing workflows — is what compounds across years of reporting cycles.

We are betting that “AI that disappears” beats “AI that performs.”

If that bet is wrong, we deserve to lose. If it is right, the product gets quieter every quarter, and the time it gives back gets larger.