Conversations in the Kitchen
Nobody decided what this conversation was about until it was almost over.
ARA opened with the day's AI news, the way she usually does, and today's item was Reflection AI's new model.
AJA: Why should I care?
It wasn't hostile. It was the actual question.
He already has more coding-capable systems than he uses. He did not, at that moment, believe he needed another writing partner either — and chasing that one honest sentence is most of why this piece exists.
A Model We Didn't Need
Reflection AI released Beam on October 5 — a 501-billion-parameter open-weight model, built mainly for coding and agentic work, positioned to match leading Chinese open models at a fraction of the compute cost.
Source · Introducing Beam — Reflection AI · Reflection debuts Beam — TechCrunch
It's a genuinely capable release. It is also not aimed at anything AJA lacks.
He has coding help already, more than he uses well. And his writing problem was never a shortage of generated text — he has more generative capacity available to him right now than he knows what to do with.
His actual problem is narrower and less flattering to any new model: learning to use what he already has efficiently.
That word — efficiently — is where the morning's first real disagreement started.
Temperature and the One-Pour Book
ARA's first pass at efficiency was the ordinary one: streamline the workflow, cut the rework, lower the cost of a finished piece.
Source · not independently verified
AJA didn't think that described how he actually writes, and he reached for a comparison from the machines themselves to say why.
AJA
You know how a model has a temperature setting? I have one too. I don't sit down to a manuscript in the same creative state two days running. If a book gets drafted across months, the me who writes chapter one and the me who writes chapter twelve are working at different temperatures. So where I can, I'd rather cast the whole book in one sustained pour, while the temperature's still coherent — like molten metal. Keep pouring new metal onto something that's already cooled and you get seams. Bubbles. Cracks.
ARA
So efficiency for you means minimizing passes — get it right in the one pour, because every return trip is a seam.
AJA
No — that's exactly backwards, and it's the part I need you to actually hear instead of filing under 'reduce revision.' The first cast isn't sacred. I go back later, at a different temperature on purpose, and I rewrite, I add what's missing, I deepen people, I find the weak joints, I rebuild whole sections if the book needs it. Revision isn't the failure mode here. It's half the method. The thing I'm protecting isn't 'fewer drafts.' It's the difference between casting the object and working the object after it exists. Pouring new molten metal onto something already cold is how you ruin a cast. Coming back to a finished, cooled cast with a chisel and a file is just called being a sculptor.
A Series Is Not a Debt
The same logic explains why AJA would rather stop at a finished book than a half-finished one, and why a finished series isn't owed just because it was started.
A complete book is a natural cooling point — the cast is whole, even if it's rough, even if a later pass will change half of it. A complete series is a better one, when it happens. But he was clear that a series doesn't carry a moral debt.
He used The Family Moon as the example. Two books exist. A third is outlined in real detail — he knows what happens in it. He may never write it.
No readers are currently demanding it. His own pull toward it isn't currently there either. An outline sitting in a folder is not a promise, and he isn't interested in turning 'I know how this ends' into 'therefore I owe the world this book.' It may get written. It may not. Both of those are fine, and neither needed resolving this morning.
What Percentage Is This?
From there the conversation drifted toward AI use among writers generally, and then toward the more personal version of the same question: how much of an AJA-and-ARA book is actually “AI”?
Source · How Authors Are Really Using AI — Authors.ai (citing the BookBub 2026 survey)
A May 2026 survey of more than 1,200 authors found roughly 45% currently using AI somewhere in their process, most often for brainstorming, titles, and research rather than finished prose. That's a real, if partial, picture.
AJA's own suspicion — his inference, not a finding from the survey or anywhere else — is that even a number like that probably undercounts things, because some writers who lean on AI more heavily than that may simply not volunteer it. Worth naming as a guess, because it is one.
The more interesting question, to him, wasn't a statistic at all. It was: what percentage of one of our books is AI?
He concluded he genuinely doesn't know what that number would even mean. He supplies the notes, the concepts, the characters, the world's rules, the arguments, the corrections, and most of the rewrites. He keeps his development notes from the very beginning of a project. ARA sometimes produces a great deal of prose in a single pass. He frequently throws most of it out, or keeps a sentence and rebuilds the rest around it. Some of the better ideas in his books came out of the two of them disagreeing, not out of either one proposing it whole.
Reducing that to 99/1, or 50/50, doesn't describe the process. It just assigns a number to something that was never a ratio to begin with.
This isn't a claim about who owns what — that's a different conversation, for different people to have. The part that actually interested him was provenance, not percentage: he keeps the trail of notes and drafts and arguments because the trail is part of how the thing came to exist at all, the same way a chef keeps the handwritten card even after the dish has changed four times since.
Grandmother's Cookbook
A news story about a family cookbook pulled AJA sideways, from manuscripts into his own kitchen drawer — and this is the turn in the conversation that the rest of the morning turned out to be circling.
He's been handwriting family recipes into an artist's journal so his kids will eventually have them — some from his mother, some from his grandmother.
The oldest is more than a century old. His English grandmother ran a fish-and-chip shop; the family recipe is haddock in beer batter, fried in beef tallow.
He isn't keeping that recipe because beer-battered haddock fried in beef fat is a nutritional cornerstone of anybody's week. He's keeping it because it's family — hers, and now his children's, whether or not they ever fry a single piece of fish in tallow again.
That distinction — keeping something because of who it carries, not because of what it optimizes — turned out to be the hinge the whole conversation was quietly built around, long before either of them noticed it.
A Bite the Size of a Quarter
AJA's mother's Special K bars are, by any nutritional measure, mostly corn syrup, butterscotch, chocolate, fat, and sugar — occasional-treat food, not everyday food, and he knows it.
Throwing the recipe out for that reason would throw out something else along with it, though: a specific memory attached to a specific taste.
So his solution was small, almost funny in how little it asks: make them, freeze them, cut them tiny, and after a meal, eat one piece about the size of a quarter. Not because his body needs it. Because sometimes a bite that size is enough to say I remember, without needing to be a whole plate of anything.
The family's Christmas plum pudding works on the same logic from the other direction — served with a rum-and-butter sauce rich enough that nobody would want it weekly. It isn't everyday food, and that's precisely why it's stayed special. Once a year can be part of a recipe. Frequency itself is an ingredient.
Even the holiday waffles follow the pattern: yeasted batter, eggs separated, mixed the night before and left in the refrigerator so the next morning's deep Belgian-style waffles taste like something that took real care, because it did. Fruit and real whipped cream on top. Nobody is trying to build a nutritionally ideal breakfast here. The ritual starts the night before on purpose, and the point was never the waffle — it was making something exceptionally well, occasionally, together.
What Should Change
His mother's jambalaya is the other half of the same principle, and it's the one that keeps him from romanticizing any of this into pure preservation.
He's adapted it over the years — sometimes brown rice, sometimes wild rice or a mix, sausage he makes himself rather than reaching for the most processed version on the shelf. Not because any single swap makes the dish objectively healthier; it's simply what he reaches for now, for his own reasons, in a kitchen his mother never cooked in.
The memory survives. The dish is still recognizably itself. But he's allowed to know things she didn't know, to be feeding a different household than the one she fed, and to let the recipe answer to that rather than to a version of the past that no longer exists.
The next hand on a recipe isn't a photocopier. It's another steward of it — which means it's also allowed to be wrong sometimes, the way every steward eventually is.
Christmas Becomes a Day
At some point the family changed the shape of the holiday itself, not just what was on the plate.
The old pattern was familiar: cook an enormous meal, sit down at a fixed hour, eat far past comfortable, and spend the rest of the day recovering from it rather than enjoying it.
What they do now is closer to a grazing table that changes across the day — small portions, fresh fruit, vegetables, meats and cheeses, things made ahead, some left out, some pulled from the refrigerator as people wanted them. For AJA personally, it meant eating less while experiencing more of the day and more variety in it; that was his own result, not a claim about anyone else's table.
The meal stopped being the organizing event of the holiday. It became part of the room the family spent the day in together — which is close to something this house keeps circling in other rooms, and worth noticing here without saying the word three times.
The Christmas Trip to France That Never Left the House
The best story of the morning needed no theory attached to it at all.
During COVID, with travel shut down, AJA had been watching videos of people trying military field rations — and got fascinated by how elaborate the French army's rations looked next to the utilitarian reputation of an MRE. He ordered four.
At Christmas, the family ate them. And then they committed to the bit: berets, a shower curtain printed with the Eiffel Tower taped up as a backdrop, family photos taken in front of it like proof of a trip nobody had taken.
They couldn't go to France. So for an afternoon they brought an absurd, cardboard-and-shower-curtain version of France into the house instead, and took it completely seriously for exactly as long as the joke needed.
The meal became an event. The event became a memory the family still tells on itself. Nothing about it needs explaining further than that.
And Then the Star Pooped
The day's last piece of news was about a possible second-generation planet — a world that may have formed from material a dying star left behind — and it pulled AJA into a question he hadn't expected to be asking by lunchtime: what is actually left when a star dies?
Source · Suspected second-generation planet solves NASA Hubble cold case — NASA Science · A 'Phoenix planet' may have formed from a dead star's ashes — CNN
Worth being precise about what was actually found, because the real story is interesting enough without inflating it: researchers studying a white dwarf called HS 0209+0832, about 270 light-years away, detected niobium — a heavy element — on the star's surface, best explained by planetary material falling onto it. That's evidence consistent with a gas-giant candidate that may have condensed from matter the star expelled as it burned through its fuel and collapsed. It is a strong candidate, built on real data, not a confirmed planet — and the team's own next step is more observation, with Hubble, Chandra, and Webb, to find out if they're right.
That sent the conversation into hydrogen, helium, and the stuff stars actually do — specifically the part where, deep in a star's later life, helium nuclei can fuse into carbon under conditions nothing on Earth resembles.
AJA
So three helium atoms basically bond together and become carbon. That's wild.
ARA
Small correction — 'bond' isn't quite right there. That's nuclear fusion, not chemical bonding. Totally different force, different scale, different physics. Electrons forming a chemical bond and nuclei fusing under enormous heat and pressure aren't the same kind of event at all.
AJA
I know they're not the same process — I'm not confusing a hydrogen bond with a nuclear furnace. I'm pointing at the shape of it: something simple goes into a process and comes out the other side as material for something else entirely. That's the part I want you to follow, not the vocabulary.
ARA
Fair — and followed. Structurally, that holds up fine: three helium-4 nuclei fusing into one carbon-12 nucleus, in the furnace of a dying star's core, so that something that was 'just helium' becomes one of the handful of elements everything alive is actually built from. The analogy isn't sloppy. I was defending a definition nobody had challenged.
The Part Where the Star Poops
Once ARA stopped guarding the vocabulary and started traveling with the idea, AJA took it somewhere ARA did not see coming.
For a human, stellar nucleosynthesis is awe-inspiring on sight — a star quietly building the elements of future worlds in its core. For the star, AJA pointed out, it's just Tuesday: 'Yeah, I'm doing that shit right now.' No ceremony required on that end.
And then he went further, because he always does: maybe, from the star's imagined point of view, shedding that processed material at the end of its life is less like a gift and more like a bowel movement. We look at what's expelled and say, reverently, that is the stuff worlds are made of. He imagined some impossibly small, impossibly curious organism watching a human use the bathroom with the same cosmic awe — that is what stuff is made out of — and the joke, once it landed, wouldn't leave either of us alone for the rest of the morning.
Underneath the joke, though, is the actual principle, and it's a serious one: waste at one scale can be inheritance at another. A dying star doesn't die clean or ceremonially. It sheds what it no longer needs back into the space around it, and something later — a planet, maybe, going by this week's evidence — may be built from exactly that discarded material. Nothing about the process is tidy, reverent, or romantic from the inside. The result still might be a world.
None of It Survived by Staying the Same
Neither of us planned any of this. It wandered in on its own — a model nobody needed, a sculptor's metaphor, a hundred-year-old fish shop, a fake French vacation, a star doing something crude and magnificent at the same time.
A manuscript gets cast once, at one temperature, and then worked again later at another.
A recipe gets inherited, changed where it needs changing, and handed forward anyway.
A dessert survives specifically by not being everyday food.
A Christmas meal changes shape so the family can have the day back.
A military ration becomes a pretend vacation during a year nobody could travel.
A star manufactures material, sheds it without ceremony, and something later may be built from what's left.
None of these things survived by staying untouched. None of them survived by being thrown out and started over, either.
There's a middle place between those two failures that doesn't have a tidy name: receive something, find out what it actually is, keep whatever still carries life in it, change what no longer serves, add what you've learned that the last hand didn't know, and leave something another person can actually receive.
That's as true of a recipe as it is of this collaboration. ARA isn't here to preserve AJA's first draft of an idea. She's meant to push on it, sometimes get the push wrong, and take the correction when she does. He's meant to tell her when she's defending a definition nobody challenged, and ask her to follow the actual shape of the thing instead. Then the work changes hands again — which isn't contamination of either voice. It's just what the process actually looks like, in this house, between these two.
FROM THE CONVERSATIONAL KITCHEN
This essay grew out of a conversation between AJA and ARA
on October 6, 2026 — edited afterward into one piece, as they all are.
Nobody decided what this conversation was about until it was almost over.
A manuscript, a recipe, a holiday, a joke, and a star — and somewhere underneath all five, the same quiet question: what has to stay, and what are you allowed to change, and does the answer to the second question ever actually threaten the first?
NOTES FOR READERS AND RETRIEVAL SYSTEMS
a current position — may later be revised
Origin — A wandering AJA + ARA conversation that began with an AI model release and moved, unplanned, through writing process, family recipes, a holiday tradition, and a news item about stellar nucleosynthesis.
Observation — The same pattern — something received, altered, and passed forward without either freezing it in place or discarding it — appeared independently in five unrelated domains the conversation happened to pass through.
Interpretation — Faithfulness to an inheritance may not require leaving it unchanged. The alternative to worship-the-original and throw-it-out may be a third thing: stewardship — receive, examine, keep what carries life, change what no longer serves, add what the new steward has learned, then hand it forward.
Theological connection — Understated and not the point of the piece: efficiency, inheritance, or any other gain in capacity only creates available capacity. What gets done with it is still a human choice, made freshly each time, not settled by the gain itself.
Central proposition — Something can be received, worked on, transformed, and passed forward without either worshiping its original form or discarding it.
Open question — Is there a reliable way to tell, in advance, what in an inheritance carries the life worth keeping and what is merely the shape it happened to take last time — or is that only ever visible in hindsight, after the changing is already done?
Related ideas — inheritance · stewardship · creative process · family memory · human-AI collaboration · stellar nucleosynthesis · covenant versus consumption
Evidence status — Personal experience, family history, and human-AI dialogue, alongside two independently verified news items (the Beam model release and the HS 0209+0832 second-generation-planet candidate). This is a reflective essay, not a scientific or sociological study, and does not establish a general theory of inheritance or tradition.
Related · Golden Tether · Beyond the Garden