On the Record · NicNac Systems
Submitted, On the Record: Our MIT Solve Application
An application is a claim until it is filed. Ours is filed: Solution 126929, submitted to MIT Solve on November 5, 2026. This post says plainly what we entered, and what we did not.
The short answer: NicNac Systems is not an AI company in the usual sense. We are a human augmentation company — we build AI that refuses to replace people and instead makes them more capable. That is the company we entered at MIT Solve, in the AI for Humanity purpose, with a founder video of 1 minute 56 seconds and the measured work on this blog standing behind every sentence of the application.
What was actually submitted
The record is specific, and we keep it that way. Solution 126929 was submitted on November 5, 2026. The application carries a verified founder video, and Bette Saffran stands as the application's backup contact. Nothing in it claims a result MIT has not announced: submitted is submitted. It is not a selection, not a finalist standing, and not a prize — and you will never read it described as one here, because the discipline that runs this blog runs the application too.
Why this door, for this company
Solve's AI for Humanity purpose asks what artificial intelligence is for. Our answer is the founding sentence of this company: augmentation, not replacement. Every system we run — the phone answered in the business's own name, the follow-up that stops slipping, the church guest who gets a personal note inside 48 hours — is built so a person becomes more capable, not less necessary. The application simply carried that sentence into a room where the question is being asked at scale.
The proof was already public
We did not write new claims for the application. The parser study on this blog — 21 of 21 checks against the paid service's 20 of 21, with the three places the paid service still beats us published in the same article — was public before the application went in. So was the AI-visibility test we ran on ourselves, in which we were recommended in 0 of 28 scored answers, published whole. An application that can point at scores a stranger can check is a different thing from an application that asks to be believed. Proof before pitch, again — this time, proof before application.
Whatever MIT decides, the work continues on the same terms: measured, published, and aimed at people. This is our time — and the record is open for anyone to read.
Read the proof
The measured studies the application stands on are on this blog, published whole.
Start with the proofOne list, measured notes only
New articles when they ship - for owners who would rather read scores than slogans.
Double opt-in - a confirmation email completes your signup. Unsubscribe any time; our suppression list is global and permanent.