The Grit

The why of the Whetstone

This one assumes you have read The Whetstone, which sets out what these tools are used for here and what they are not. That document says what happens. This one says why, and it exists because somebody asked.

Why disclose at all

Nobody requires it. There is no law, no publisher, and no store demanding that a small independent house account for how its books get written.

I disclose because the alternative is that somebody finds out later. A person who buys a book, likes it, and discovers afterwards that something was withheld has been handled rather than met. That is a worse outcome than losing the sale in the first place.

And because I ask it of the people I work with. Collaborators are asked not to hand back generated work in place of their own hand, and a house that asks that while staying quiet about its own practice has no standing to ask anything.

What the tool actually does here

The short version in The Whetstone says drafting and checking. Neither word describes what the work feels like.

I frame something, usually at length. Something comes back. Most of the time the first draft gets torn apart: the characters are wrong, a mechanic that sounded fine is nonsense on second reading, an entire dimension is missing and the draft does not know it is missing, because a draft cannot know what is absent.

Then I say what is wrong and it comes back again, and somewhere in that loop the thing appears.

The value is not the drafting. It is having something specific to argue with. Reading what came back and seeing immediately what it is not is faster than generating from nothing, and it is a different mental act. A blank page asks what do you want. A wrong draft asks what is wrong with this, and I can answer the second one instantly.

So the tool is closer to a sparring partner than to a writer. It is there to be corrected, and the correcting is where the work happens.

What it gets wrong

Even after months of working with these tools, and building a detailed set of rules to prevent the errors I had already found, they still occur. These six all happened within a few weeks of each other.

•  It miscounted a scale I designed, one the whole game line rests on, and did it twice, then wrote the error into a scenario as though it were settled.

•  It tried to sweep a mechanic out of a book as a leftover from an old draft. It was not a leftover. I had folded it into another system a year earlier and nobody but me knew that.

•  It made the same rules mistake in two different books, treating the presence of an enormous creature as damage in itself. That is not how the system works. I sent it back to read the rule, twice.

•  It wrote that a character dies first in a battle. The dice had not been rolled.

•  It let a word from the spy book leak into the superhero book, flattening two registers meant to sound nothing alike.

•  It built a substantial piece of work on the assumption that two abilities were one, when I had written them as separate years earlier in a different book. The whole sequence had to be redone.

None of those are grammar. Each was a judgement about material I had built over years, and each was invisible to the thing that produced it.

That is the reason the tool cannot be a primary creative force here. Not that it writes badly, because it often writes fluently. It cannot see what a thing means, so it cannot tell which parts carry the meaning and which are scaffolding. It will remove the load-bearing wall and leave the decoration, and it will do it confidently.

There is also a kind of material these tools produce that a reader can feel. Details that contradict each other. Facts that are wrong. A thin, rushed quality where the sentences are fluent and the substance is not there. Nobody has to be told what has happened; they simply notice that something is off.

The noticing is the problem. Once somebody is looking at a thing and finding faults in it, they are outside the experience rather than inside it, and getting them back is hard. That is the same failure as the six above, at a larger scale: the tool cannot see which parts carry the meaning, so it cannot tell when it has produced fluent nothing.

The creativity is still ours. We see things with meaning that these tools cannot.

Where the ideas come from

Anybody can claim the ideas were theirs, so here is what is actually underneath.

Alex Raymond and the shape of a hero walking into a tyrant’s court. Kirby, and gods who are machines. Zelazny, and a pantheon that turns out to be a family with a grudge. Serling, and the turn that is honest rather than clever. Welles, and the pleasure of a thing that is gloriously not what it appears to be. Mythology from a dozen places. Kung fu cinema. Kaiju films. Spy fiction of the cold and disappointed sort. Stage magic, which I practise, and which taught me to build everything backwards from the moment somebody feels something.

Decades of loving things and wondering why they worked, and it comes out as collision: a genre thrown against a genre it has no business meeting, then held there until something true falls out.

That collision happens before any tool is opened. It is not promptable, because it comes out of what one person happened to read and watch and do, in what order, over a lifetime.

Why homage gets named

Where a work is built on somebody else’s, that is named, and this predates any use of these tools.

The reason is simple. If something is in there, and I know it is in there, then not saying so is a small theft. Naming it costs nothing and it is what I would want if it were mine. The full account of what is owed to the work a thing came from is Keep the Bones.

What authorship actually rests on

Not the implement. The judgement, exercised over time, and the willingness to be responsible for the result.

Everything that ships does so because I decided it was worth keeping, or rewrote it, or tore it down and built it again. When something in one of these books is wrong, it is wrong because I did not catch it. There is nobody else to point at, and that is the whole of it.

A stone wears down and the blade does not. The stone gives itself up to the work and the edge is what remains.

Where this document ends

This is the reasoning as it stands. It will change as the practice changes, and this document changes with it.

Bruce is the author. AI is the whetstone.