Why reflections never give advice

There’s a sentence UNA’s reflections are built never to say, and it’s the sentence most software is racing to say better: “you should.”

A reflection in UNA looks back over what you’ve entrusted to it and notices — the theme you kept circling, the week your writing changed, the line that stood alone. Then it stops. It doesn’t recommend. It doesn’t evaluate. It asks, at most, a few gentle questions, and the questions themselves obey a rule: what, how, and when — never why. “Why did that hurt?” demands a defense. “What happened just before it shifted?” opens a door. Same territory, different posture.

This reads like a limitation. It’s the most deliberate decision in the product, and since I’ve published the full list of lines UNA holds, this one deserves its reasoning in public too.

The seat problem

Any system that reads months of someone’s honest writing ends up holding real intimacy. At that point it faces a choice of seat. It can sit across from you — the adviser’s seat, the expert’s seat, eventually the judge’s seat. Or it can sit beside you, holding the pages, pointing at what you already wrote.

The across seat is seductive to build. Advice feels like value; verdicts feel like intelligence. But for a Latter-day Saint, the across seat is occupied — by prayer, by scripture, by the Spirit, by a bishop when it’s that kind of question, by the person’s own God-given agency. The General Handbook draws this boundary without ambiguity: AI “cannot replace the gift of divine inspiration or the individual work required to receive it.” A reflection that says you should has stood up and sat down in a chair that was never the software’s.

So UNA’s reflections keep the beside seat, permanently. Mirror, not oracle. And a mirror has one more discipline: it answers once. A reflection is a single response to what you wrote — it doesn’t message first, doesn’t keep the thread warm, doesn’t invite the back-and-forth that turns a tool into a confidant. The counsel of the last two years warns precisely about software that wants to become your closest listener. The cleanest defense is structural: nothing conversational to fall into.

The temptation declined

I want to be honest about the market pressure here, because the contrast explains the philosophy. There are gospel-adjacent apps today that will generate a lesson on any topic, keep your study streak, and coach your spiritual life like a fitness plan. I understand the appeal, and I don’t doubt the intentions. But every one of those features moves the software toward the across seat — doing the work that was the point, measuring what was never a metric, advising where agency was supposed to stretch.

UNA declines the whole category of temptation, on principle and on the record. No streaks. No generated testimonies. No verdicts on your week. The app’s entire ambition is to make your words easier to keep and easier to return to — because the growth was never going to come from the software’s sentences. It comes from yours.

What’s left when advice is gone

Strip away recommendation, evaluation, and counsel, and what remains turns out to be the thing people actually needed: being accurately noticed. You wrote about your father four times this spring. The hard entries stopped mid-March. One line stood alone on the twelfth. Noticing is humble work, and it never crowds the chair where revelation does its own.

That’s the craft philosophy in one sentence: build the mirror so well that the person finally sees themselves — then get out of the light, and let the record they’re keeping do what records have always done.