Prompting
Muse
A second opinion that reads your board and suggests a move, and stays quiet when it has nothing worth saying.
Muse reads your board and sometimes says nothing. The silence is deliberate, and it is what keeps the suggestions worth reading when they do arrive. ## How it decides Two stages, and they have different jobs. The first runs hot and wide. It takes an inventory of what is on your board and generates several different candidate moves, deliberately diverse, because the first idea about a board is rarely the interesting one. The second runs cold and careful. It reads those candidates against what is actually on the canvas, picks the one that is grounded in what it can see, and then asks a final question: **is this worth saying at all?** If the answer is no, Muse stays quiet. :::note[Why the silence gate matters] An assistant that comments on everything gets ignored within a day, and then it is noise you have to look past to work. One that speaks a handful of times in an afternoon gets read. ::: ## What a suggestion looks like A note on the canvas, naming a specific move: a tool to try on something already on the board, with a reason drawn from what is there. Suggestions are grounded rather than generic. Muse can see your assets, so it proposes against them instead of offering advice that would fit any board. ## Where it helps **Naming the tool you forgot.** There are more of them than anyone holds in their head. The most common useful suggestion is a tool that fits exactly and that you had not thought of. **Breaking a stall.** When a board has gone round the same idea four times, an outside read is worth more than a fifth attempt. **Noticing what you have.** A reference you dropped an hour ago and stopped seeing is still on the board, and Muse is reading the board rather than your memory of it. ## Its limits It has no taste of its own and it is not an art director. It reads what is present and suggests a plausible next move, which is a different thing from knowing what the work should become. Treat a suggestion the way you would treat a colleague leaning over your shoulder: occasionally exactly right, occasionally worth ignoring, and never the deciding vote. :::tip[Teach it what you like] Muse reads a taste profile when you have set one, so its suggestions narrow toward the kind of work you actually make. Setting that up is covered in [your creative preferences](/docs/using-plnty/prompting/your-creative-preferences/). :::Muse reads your board and sometimes says nothing. The silence is deliberate, and it is what keeps the suggestions worth reading when they do arrive.
How it decides
Section titled “How it decides”Two stages, and they have different jobs.
The first runs hot and wide. It takes an inventory of what is on your board and generates several different candidate moves, deliberately diverse, because the first idea about a board is rarely the interesting one.
The second runs cold and careful. It reads those candidates against what is actually on the canvas, picks the one that is grounded in what it can see, and then asks a final question: is this worth saying at all?
If the answer is no, Muse stays quiet.
What a suggestion looks like
Section titled “What a suggestion looks like”A note on the canvas, naming a specific move: a tool to try on something already on the board, with a reason drawn from what is there.
Suggestions are grounded rather than generic. Muse can see your assets, so it proposes against them instead of offering advice that would fit any board.
Where it helps
Section titled “Where it helps”Naming the tool you forgot. There are more of them than anyone holds in their head. The most common useful suggestion is a tool that fits exactly and that you had not thought of.
Breaking a stall. When a board has gone round the same idea four times, an outside read is worth more than a fifth attempt.
Noticing what you have. A reference you dropped an hour ago and stopped seeing is still on the board, and Muse is reading the board rather than your memory of it.
Its limits
Section titled “Its limits”It has no taste of its own and it is not an art director. It reads what is present and suggests a plausible next move, which is a different thing from knowing what the work should become.
Treat a suggestion the way you would treat a colleague leaning over your shoulder: occasionally exactly right, occasionally worth ignoring, and never the deciding vote.