That’s how long I spent writing about my life with Ellen, my wife of 47 years who died on Thanksgiving Day 2019. Five years of wrestling with memories, confronting regrets, trying to honor a complicated woman and an even more complicated marriage. Five years of starting over, deleting thousands of words, questioning whether I was smart enough or talented enough to write something she’d be proud of.

And here’s the part that’s going to make some of you uncomfortable: I didn’t do it alone. About 15% of the final manuscript was created with significant help from Claude, an AI assistant. Not ghost-written. Not generated by typing prompts into a machine. But genuinely collaborative in ways I’m still processing.

This is the story of that five-year journey, what I learned about grief and writing and perfectionism, and why inviting an AI into the most personal project of my life might have been the smartest—or strangest—decision I ever made as a writer.

The Project Ellen Would Have Wanted (And Also Hated)

Let me start by telling you about Ellen.

She was brilliant. Completed her graduate coursework in English literature with high honors. Started a dissertation on Shakespeare’s Winter’s Tale that her advisor said was original and would land her a job at a top university. That was the plan—she’d finish her PhD, get a faculty position, and I’d finally have time to write those bestselling novels I was so confident I could produce.

Except Ellen never finished her dissertation.

She was the kind of perfectionist who kept expanding the project instead of narrowing in on completion. New ideas would emerge during the writing, connections she hadn’t seen before, and suddenly the dissertation needed another chapter, another revision, another year of work. I’d stay up nights typing her papers (this was before word processors), and she’d get new insights mid-typing and we’d have to start over.

It drove me crazy. My philosophy has always been “good enough is good enough—turn it in and move on.” Hers was “keep working until it’s as perfect as your mind can make it, even if that takes forever.”

Guess Which One…

Guess which one of us never finished her dissertation? Then take a shot at which approach I had to learn to respect, even when it frustrated the hell out of me?

Now here’s the cosmic joke: when Ellen died and I decided to write a memoir about our life together, I turned into her. The book I thought would take six months became a five-year obsession. What started as a simple collection of sweet memories morphed into something increasingly complicated—part tribute, part honest reckoning with a complicated marriage, part meditation on grief and regret and the meaning of a life shared.

I often felt like I was wrestling the manuscript in a tub of cold jello. I wrote thousands and thousands of words that got deleted. The project kept expanding instead of narrowing. I couldn’t figure out how to frame it, where to put the focus, what story I was actually trying to tell.

I’d become the perfectionist I used to be impatient with. Funny how that works.

Tolkien, Niggle, and the Purgatory of Creation

Four years into the project, still stuck, I just happened to discover Tolkien’s short story “Leaf by Niggle.”

If you don’t know it, here’s the relevant part: Niggle is a painter obsessed with capturing a single tree on canvas, getting every leaf perfect, constantly interrupted by mundane obligations but unable to let the painting go. He dies before finishing it. The story’s second half takes place in a kind of purgatory where Niggle finally sees his tree—his incomplete, imperfect tree—made real and whole in ways he never imagined possible on earth.

The story gutted me.

Because that’s exactly what I was doing. Obsessing over getting Ellen’s memoir perfect. Trying to capture every leaf on this impossible tree I was painting. Spending years in a kind of creative purgatory, revisiting memories both beautiful and painful, confronting things done and not done properly, questioning whether I’d ever finish or whether I’d die with this manuscript incomplete like Ellen’s dissertation.

The parallel to my own experience was eerie. Tolkien wrote “Leaf by Niggle” after a serious illness when he feared he might die before finishing The Lord of the Rings. I’m 77 now. I started this memoir at 72. Death isn’t an abstract concept anymore—it’s a neighbor I see every morning when I walk past the houses of the other widowers in my neighborhood.

Would I finish Ellen’s memoir before I joined her? Would I get to see my tree completed, or would this project remain forever unfinished like her dissertation?

When AI Became My Collaboration Partner

Here’s where the story gets interesting, and where I suspect I’ll lose some of you.

Almost five years into the project, I had a compelling idea: adding sections to the memoir written from the point of view of both living and dead writers, poets, philosophers, even scientists might finally give me the hook I needed to excavate the depths of grief that I didn’t have the smarts or the skill to mine. Since I didn’t have the skills or the detailed knowledge, I decided to turn to Anthropic’s Claude Sonnet AI to see if it could write drafts of these imagined sections for me to review and then either use or discard.

To clarify, the memoir needed supporting material—contextual pieces that would help readers understand grief more broadly, that would place my personal experience within the larger landscape of human loss. I’m a thriller writer by training. I can plot a creepy story. I can write snappy dialogue. But crafting imaginary journal entries from Hemingway or Mark Twain grappling with grief? Writing historical letters of condolence in the voice of Spinoza? Creating poems about loss that felt authentic but weren’t mine?

When You Need Help…

That’s not in my skill set.

So I asked Claude to help. And what emerged was something I hadn’t anticipated—a genuine creative collaboration.

I’d provide the concept: “I need a journal entry from Jung reflecting on the death of his wife.” Claude would create a draft. I’d read it, feel whether it worked emotionally in the context of my narrative, ask for revisions, push back on phrasings that felt wrong. We’d go back and forth until we had something that served the larger purpose of the memoir—enriching my personal story with broader perspectives on grief and loss.

The result? About 15% of the final manuscript consists of these supporting materials—journal entries, letters, poems, brief essays—all created through this collaborative process.

The Authenticity Question (And Why It’s Complicated)

I know what some of you are thinking. “That’s not real writing. That’s cheating. How can you claim this memoir is authentic if a machine wrote part of it?”

Fair question. Let me complicate it for you.

Ellen’s memoir is the most emotionally authentic thing I’ve ever written. Every memory, every moment of grief, every painful recognition of my failures as a caregiver—that’s all me, straight from the heart, no AI assistance. The core narrative is 100% human.

But here’s the thing: I wanted the memoir to be more than just my personal story. I wanted it to be useful to other grievers. I wanted it to place my specific loss within the broader context of how humans have always grappled with death and grief. That required perspectives and voices I couldn’t provide on my own.

Could I have spent years learning to write convincing historical letters and journal entries? Sure. Would that have made the book more “authentic” somehow? I honestly don’t know.

What I do know is this: Claude’s contributions helped me create a richer, more textured memoir than I could have created alone. The imaginary Hemingway journal entry helped me articulate aspects of masculine grief I was struggling to express. The Spinoza letter gave me language for philosophical dimensions of loss that were beyond my ability to capture. The poems about grief created emotional resonance at key moments.

Were these “real”? They felt real to me. They served the truth I was trying to tell, even if they weren’t created by my hand alone.

What “Good Enough” Actually Means

Remember Ellen’s perfectionism? Remember how she could never finish because there was always one more revision, one more insight, one more connection to explore?

Working with Claude helped me understand something about perfectionism that I’d never grasped when I was watching Ellen struggle with her dissertation: sometimes perfectionism isn’t about making something perfect. It’s about fear. Fear that if you call it finished, it might not be good enough or fear that people will see your work and find it lacking. Fear that you’re not smart enough, talented enough, worthy enough to produce something of value.

I carried all those fears into Ellen’s memoir. I’m not a literary writer. I’m a guy who wrote paperback thrillers. Ellen always wanted me to write something serious, something with literary merit. When she died, this memoir became my chance to write the serious book she’d always wished I would write.

Am I Smart Enough?

But what if I wasn’t smart enough? What if I couldn’t write at the level a memoir of our life together deserved?

Here’s what the five-year journey taught me: “good enough” doesn’t mean settling for mediocrity. It means recognizing when additional revision starts being about fear rather than improvement. It means trusting that the work you’ve done—imperfect as it is—has value.

The joy I finally felt when I could see the memoir’s completion wasn’t about achieving perfection. It was about reaching a place where I could say: “This is what I have to offer. This is my tree, painted with whatever skill I possess. It’s finished.”

And here’s the profound part: that satisfaction transcends any concern about whether the memoir finds readers, whether it gets reviewed, whether it leaves a legacy. Doing the work itself—wrestling those five years with memory and grief and language—that’s what mattered. The transformation it created in me is the real achievement.

The Technical Questions about Writing a Grief Memoir

Okay, let’s get practical. If you’re curious about AI collaboration for your own memoir or writing project, here’s what I learned:

• What worked: Using AI for supporting material that enriched the main narrative. Claude created content that required skills I don’t have—historical voice, poetic language, philosophical reflection. This freed me to focus on what I do well: straightforward narrative, emotional honesty, storytelling.
  
• What didn’t work: Any attempt to have AI write my personal memories. I tried once, early on, describing a memory to Claude and asking for a draft. The result was technically competent but emotionally dead. My memories have to be in my words or they’re lies.
  
• The transparency question: I was clear from the beginning that I’d acknowledge AI’s contribution. The memoir’s foreword explicitly states that less than 15% was created with an AI language model, and I explain exactly how that collaboration worked. I’m not trying to pass off AI-generated content as entirely my own work. That would be dishonest.
  
• The control question: Every piece of AI-generated content went through multiple revisions based on my direction. I made all final creative decisions. If something didn’t serve the memoir’s larger purpose, it got cut regardless of how well-crafted it was. The AI was a tool, not a co-author in the traditional sense.
  
• The future question: I genuinely don’t know if AI collaboration will become standard in memoir writing or if it’ll be seen as a weird experimental phase we look back on with embarrassment. What I do know is that for this particular project, at this particular moment in my life, it helped me create something I couldn’t have created alone. On my 77th birthday—January 2025, five years after starting this project—I wrote the final words of Ellen’s memoir.

Surprised by the Ending…

The ending surprised me. I’d planned to conclude with fetching her cremated remains from the funeral home. But when I brought her home and set the cardboard box on my desk, fragments of a poem started appearing. Not from me, exactly, but from that mysterious place where language sometimes emerges without conscious effort.

That poem became the actual ending. Unexpected. Imperfect. But true.

I thought about Niggle again. About how in Tolkien’s story, the painter finally gets to see his tree—the one he’d labored over his whole life, never finishing, never satisfied—made real and complete in ways he couldn’t have imagined when he was obsessing over individual leaves.

That’s what happened with Ellen’s memoir. I’d spent five years wrestling with it, never quite satisfied, always seeing ways it could be better. But when I wrote that final poem and knew I was done, I experienced something like what Niggle must have felt.

The memoir isn’t perfect. It’s a 77-year-old thriller writer’s attempt to capture a complicated 47-year marriage and process his grief. It includes imaginary journal entries and AI-assisted poems alongside my own memories and reflections. It’s messy and experimental and probably breaks several unwritten rules about what memoirs should be.

But it’s finished. It’s true, as true as I could make it. And it exists now, independent of whether anyone reads it.

That’s enough.

What This Means for Writing a Grief Memoir

If you’re working on your own memoir—whether about grief or any other significant life experience—here’s what I learned that might help:

• Let go of perfectionism. Not by settling for mediocre work, but by recognizing when you’re revising out of fear rather than genuine improvement. Ellen never finished her dissertation because she couldn’t call anything done. Don’t let that be you.
  
• Consider all your tools. AI collaboration isn’t for everyone, and it shouldn’t be. But if there are aspects of your memoir that require skills you don’t have, consider whether technology might help you achieve your vision. Just be transparent about it.
  
• Trust your emotional truth. The parts of your memoir that come from direct personal experience need to be in your voice, with your words. That’s non-negotiable. But supporting material, contextual elements, things that enrich the narrative without being the narrative—there’s more flexibility there.
  
• Remember why you’re writing. On the hard days—and there will be many—reconnect with your purpose. For me it was honoring Ellen and helping other grievers. What’s yours? Let that guide your decisions.
  
• Define success on your own terms. My memoir will probably sell a few hundred copies to family and friends and random readers who stumble across it. That’s fine. The real success was completing it, processing my grief through writing, and creating something Ellen would recognize as serious work. Your success metrics will be different. That’s how it should be.

The Tree Is Real Now

There’s a line from Leonard Cohen that I included near the end of Ellen’s memoir: “Poetry is just the evidence of life. If your life is burning well, poetry is just the ash.”

Five years of writing Ellen’s memoir was evidence of a life burning. The grief burned. The love burned. The regret and the gratitude and the terrible recognition that I’d never fully appreciated her while she was alive—all of it burned.

The memoir is the ash. The evidence that something intense and transformative happened. That’s all any piece of writing ever is—evidence that someone was here, felt these things, tried to make sense of the chaos.

Whether you use AI or not, whether you’re a perfectionist like Ellen or a “good enough” person like me, whether you finish in six months or six years—if you’re writing memoir, you’re doing the same thing Niggle did. You’re trying to capture something true on the page. You’re painting your tree.

And here’s what I learned after five years: the tree becomes real not when it’s perfect, but when you finally stop working on it and let it stand.

Mine is standing now. Imperfect. Collaborative. Human and machine and memory and grief all tangled together.

But real. Finally real.

Ellen: A Memoir of Love, Life, and Grief by Chet Day is available now on Amazon Kindle. Click here to check it out.

The post What I Learned Writing a Grief Memoir About My Wife: Five Years, One AI, and the Question of What Makes Writing “Real” appeared first on Chet Day & CasaDay Press.

In my first two posts, I talked about what documentary fiction is and how structure shapes meaning. Now let’s get completely practical and discuss writing documentary fictions, the tools that help, and the ethical responsibilities you can’t dodge.

Because here’s what I wish someone had told me before I started The Healing Physicians: documentary fiction demands a different workflow than regular writing.

Try to write it like a novel or like academic history, and you’ll fail. It requires its own methodology.

Let me walk you through what actually works.

The Research Trap (And How to Avoid It)

Every documentary fiction project starts with research. Makes sense, right? You can’t write about documented facts without first discovering those facts.

But here’s the trap I fell into: I kept researching. And researching. And researching some more.

There’s always one more source to check, one more archive to explore, one more perspective to consider. Research feels productive because you’re learning, you’re gathering material, you’re being thorough.

But you’re not actually writing.

The solution isn’t to stop researching—documentary fiction needs solid research foundation. The solution is to recognize the sweet spot between under-researched and over-researched.

Research until you understand the basic arc. For The Healing Physicians, that meant:

• Basic timeline of Shelton’s arrests and major events
• Representative journal entries showing his voice and thinking
• Understanding of Natural Hygiene lineage and philosophy
• Foundation documents showing Rockefeller’s grants and strategy
• The Carlton bankruptcy case facts

I didn’t have every detail and didn’t know exact dates for most of Shelton’s 31 arrests. I hadn’t tracked down every Rockefeller grant letter.

But I had enough to start writing with [VERIFY] markers where gaps existed.

Then I started drafting. Because here’s what nobody tells you: you discover what you actually need to know through the writing process itself.

When I’m drafting Shelton’s section and I write “I couldn’t understand why they kept arresting me,” I realize I need to verify: Was there a pattern to the arrests? Were they all in Texas or did they cross state lines? Did any result in convictions?

Those questions only emerged because I was writing, not researching in the abstract.

The Research-Writing Cycle

Here’s the workflow that actually works:

1. Initial research (2-4 weeks): Understand the basic story, gather primary sources, identify key players and events
2. Draft first section (1-2 weeks): Write with [VERIFY] markers everywhere you’re uncertain
3. Targeted research (1 week): Hunt down answers to the specific questions the draft revealed
4. Revise with verified facts (1 week): Replace [VERIFY] markers with documented claims or qualified statements
5. Repeat for each section

This cycle prevents both under-researching (which creates factual errors) and over-researching (which prevents you from ever finishing).

The Tools That Actually Help

I’m going to save you months of trial and error by telling you what tools proved genuinely useful versus what sounded good but didn’t help.

Essential Tools

1. The Facts Database (Spreadsheet)

I already mentioned this in my last post, but I’m emphasizing it again because it’s that important. A simple spreadsheet where every significant claim has its documentation.

Format:

• Column 1: Event/Claim
• Column 2: Date (or “undated”)
• Column 3: Source (primary or secondary)
• Column 4: Verification Status
• Column 5: Notes/Context

This is non-negotiable. Without it, you’ll spend hours re-verifying facts you already checked, or worse, you’ll lose track of what’s verified and what’s assumption.

2. Source Document Library (Organized Files)

Keep PDFs, scanned documents, and web archives organized by source type:

• /primary_sources/rockefeller_foundation_reports/
• /primary_sources/shelton_journals/
• /secondary_sources/biographies/
• /secondary_sources/academic_papers/

Name files with dates: rockefeller_foundation_annual_report_1913.pdf

Why this matters: When you’re fact-checking and need to verify a grant amount from 1913, you can find the relevant document in 30 seconds instead of 30 minutes.

3. Timeline Document (Visual)

Create a simple timeline showing when major events happened across all your subjects:

1901 - Rockefeller creates Rockefeller Institute
1903 - General Education Board established
1910 - Flexner Report published
1922 - Shelton opens health school in Texas
1928 - First documented Shelton arrest

[etc.]

This prevents timeline errors and helps you see patterns across the story.

Tools That Sounded Good But Didn’t Help

Fancy note-taking software: I tried Roam Research and Obsidian. They created more organizational overhead than value. A spreadsheet and organized folders worked better.

Citation management software: Designed for academic papers, not narrative nonfiction. Too much friction for what you actually need (which is just: can I verify this claim?).

AI research assistants (without human verification): AI will confidently cite sources that don’t exist. It will treat the Rikers Island myth as fact because it appears online. It can help summarize documents you give it, but it cannot replace human verification.

Working with AI: What Actually Helps

Since I’m collaborating with Claude on The Healing Physicians, people constantly ask: what role does AI actually play?

Let me be brutally specific about what works and what doesn’t.

What AI Can Help With

Drafting from detailed frameworks: Once I’ve done research, verification, and structural decisions, AI can generate draft text quickly.

Example conversation with Claude:

“I need a section in Shelton’s voice (first person, confused, angry) covering his arrests between 1928-1935. Here are the verified facts from my spreadsheet: [pastes data]. Voice should be raw, questioning, using details from his journals where he wrote ‘I couldn’t understand why they kept targeting me.’ Draft 800 words.”

Claude generates a draft. I read it against my sources, verify every claim, revise heavily for voice authenticity, and end up with maybe 60% of the original draft but written in a fraction of the time it would have taken me to write from scratch.

Maintaining consistency: AI catches inconsistencies across 36,000 words. Did I call it “General Education Board” in Part 1 but “Rockefeller General Education Board” in Part 2? Did I say $45 million in one section but $40 million in another?

Processing long documents: I can give Claude a 50-page foundation report and ask “Extract all medical education grants between 1910-1915.” It summarizes quickly.

But—critical—I verify the summary against the original document before using any information.

What AI Cannot Do (And You Must)

Research: AI can’t distinguish credible sources from garbage. It will confidently cite sources that don’t exist or facts that sound plausible but aren’t documented.

Verification: AI treats “widely reported online” the same as “documented in primary sources.” It would have included the Rikers Island myth without hesitation.

Ethical choices: Should you include this detail? Should you qualify this claim? Should you present this person’s actions more sympathetically? These require human judgment about real people and real consequences.

Narrative architecture: Which perspective serves the story? How should facts be ordered? What structure reveals the pattern? This is where documentary fiction lives or dies, and AI can’t do it.

The AI Collaboration Workflow That Works

Here’s exactly how I use AI in practice:

1. Human does research and verification (100% me)
2. Human creates structure and framework (100% me)
3. Human writes detailed prompt with verified facts (100% me)
4. AI generates draft from framework (Claude)
5. Human revises heavily, verifies every claim (100% me)
6. Human makes all ethical and narrative decisions (100% me)

The ratio is probably 70% human work, 30% AI assistance. AI speeds up drafting but doesn’t replace any of the critical thinking, research, or ethical responsibility.

If you’re not comfortable verifying every single claim an AI generates, don’t use AI for documentary fiction. The risk of including unverified “facts” is too high.

The Ethical Responsibility You Can’t Dodge

Here’s what nobody says about documentary fiction but absolutely should: you’re trafficking in real lives and real consequences.

When I write that Shelton was “confused and angry,” I’m characterizing a real person who died. If I write that Rockefeller “believed he was advancing scientific progress,” I’m attributing motive to someone who can’t defend himself.

When I write that Bill Gates is replicating Rockefeller’s playbook, I’m making claims about a living person with significant power and resources.

This isn’t abstract. This is serious.

The Four Ethical Principles

1. Fairness (Even to People You’re Criticizing)

Rockefeller systematically suppressed alternative medicine. That’s documented. But I also need to show that he genuinely believed he was advancing scientific progress, that he used homeopathy himself, that his motivations were complex.

The temptation in documentary fiction is to make villains one-dimensional because it serves your narrative. Resist this. Real people contain multitudes. Show that complexity even when it complicates your argument.

2. Precision (Distinguish What You Know from What You Infer)

I can document that Rockefeller gave $180 million to orthodox medical schools and $0 to homeopathic institutions. That’s fact.

I can reasonably infer that this wasn’t accidental—the pattern is too consistent. That’s inference based on documented evidence.

I cannot claim to know his private thoughts beyond what his journals reveal. That would be speculation.

The language matters:

• Documented: “Rockefeller gave $180 million…”
• Reasonably inferred: “This pattern suggests deliberate strategy…”
• Speculation (avoid): “Rockefeller secretly hated homeopathy…”

3. Humility (Acknowledge Uncertainty When It Exists)

“Shelton was arrested approximately 31 times” is stronger than “Shelton was arrested 31 times” when you don’t have exact documentation for each arrest.

“Exact dates for most arrests remain unknown” is honest. Inventing plausible dates would be dishonest.

Readers trust precision about what you know and honesty about what you don’t. Faking certainty destroys credibility.

4. Care (These Were Real People, Not Narrative Devices)

Shelton died in 1985. Rockefeller died in 1937. But they had families, friends, communities who might read this. And Bill Gates is still alive.

Every person in your documentary fiction deserves to be portrayed as completely as evidence allows—even when that complexity makes your argument harder.

This isn’t sentimentality. It’s professional responsibility.

The Most Common Mistakes (And How to Avoid Them)

After finishing The Healing Physicians, I can see where documentary fiction projects typically fail:

Mistake 1: Trusting Secondary Sources Without Verification

Secondary sources repeat each other. One biography includes an unverified claim, others cite that biography, and suddenly it “feels” documented because it appears everywhere.

The fix: Always try to get back to primary sources. For important claims, multiple sources that don’t just cite each other.

Mistake 2: Smoothing Over Gaps with “Reasonable” Assumptions

You know something happened, you just don’t know exactly when or how. So you fill in the gap with what “probably” happened based on similar situations.

This is fiction masquerading as documentary.

The fix: Acknowledge the gap. “Exact details of the arrest remain undocumented” is better than inventing plausible details.

Mistake 3: Letting Structure Override Truth

You’ve built a beautiful structure, and then you find a fact that doesn’t quite fit. So you massage it, or downplay it, or omit it.

The fix: Revise your structure to accommodate inconvenient facts. If the truth doesn’t fit your argument, your argument is wrong.

Mistake 4: Using AI Without Verification

AI generates plausible-sounding text with confident citations to sources that don’t exist. If you use that text without verification, you’ve published fiction as documentary.

The fix: Verify every single claim an AI generates. Treat AI drafts as unverified until you’ve checked them against your sources.

Mistake 5: Forgetting You’re Not a Lawyer

Documentary fiction lets you make strong arguments, but you’re still bound by accuracy and fairness. Make defamatory claims about living people, and you might face legal consequences.

The fix: When writing about living people, be extra careful with verification. When in doubt, qualify your language.

Your First Documentary Fiction Project: Start Here

If you’re ready to try documentary fiction, here’s your practical roadmap:

Step 1: Pick Your Subject

Choose something you’re genuinely obsessed with. You’ll spend months with this material—it needs to sustain your interest through tedious fact-checking.

Start small: One person, one decade, one transformation. Not three people across 125 years. Build skills before attempting epic scope.

Choose documented subjects: You need primary sources. If your subject left no journals, letters, or documented evidence, documentary fiction will be nearly impossible.

Step 2: Build Your Source Foundation

Before writing a single word:

• Identify available primary sources
• Gather secondary sources
• Note what’s missing (gaps you’ll need to acknowledge)
• Create your facts database spreadsheet

This might take 2-4 weeks. Don’t rush it.

Step 3: Create Your Structure

Decide:

• Whose perspective(s) will you use?
• What order will best reveal meaning?
• How many parts/sections?
• What voice for each section?

Write this out explicitly before drafting. Your structure is your argument.

Step 4: Draft with [VERIFY] Markers

Write momentum first, verification second. Mark everything uncertain with [VERIFY] so you can find it later.

Target 500-1000 words per section for your first project. Don’t overreach.

Step 5: Verify Everything

This is where it gets tedious. Every claim marked [VERIFY] needs:

• Source identification
• Verification against primary source when possible
• Multiple sources for important claims
• Qualification if certainty isn’t possible

Budget as much time for verification as for drafting.

Step 6: Revise with Verified Facts

Replace [VERIFY] markers with:

• Documented claims (with precise details)
• Qualified statements (“approximately,” “sources suggest”)
• Removal of anything you can’t verify

Step 7: Test Your Claims

For every significant claim, ask:

• Can I cite a source?
• Can I defend this if challenged?
• Am I stating fact or inference?
• Have I confused “widely reported” with “documented”?

If you can’t answer confidently, revise or remove.

What Success Actually Looks Like

Let me be honest about what “success” means in documentary fiction, because it’s not the same as fiction or journalism.

Success is NOT:

• Millions of readers (the audience for documentary fiction is smaller than for pure fiction)
• Easy writing (this is hard, tedious work)
• Fast production (verification takes as long as drafting)

Success IS:

• Every claim you make is defensible
• Readers trust that what you’ve written is true
• The narrative makes documented facts feel urgent and relevant
• Complex people are shown completely, not reduced to types
• The pattern you’ve revealed feels inevitable, not forced

The Healing Physicians succeeds (when it does) not because it reaches millions but because readers finish it and think “I can’t believe I didn’t know this” and “I can’t unsee this pattern now.”

That’s what documentary fiction can do that neither journalism nor fiction does alone.

The Question That Matters

After walking through all this process—the research workflow, the tools, the AI collaboration, the ethical responsibilities—you might be wondering: Is this worth it?

Is documentary fiction worth the months of research, the tedious verification, the ethical complexity, the smaller audience?

For me, the answer is yes. Because we need true stories that hit like novels. We need documented facts that create undeniable patterns. We need history that feels urgent and relevant to right now.

We’re drowning in information but starving for meaning. Documentary fiction provides both.

It’s harder than fiction. It’s more constrained than journalism. But when you pull it off—when you create something that’s both verifiably true and emotionally gripping—you’ve done something important.

You’ve made history matter.

And at 77, after 50 years of writing about natural health and institutional suppression, that feels like work worth doing.

Are you working on a documentary fiction project? What’s your biggest challenge right now? I’d be genuinely interested in hearing about it through the contact page. Sometimes just explaining your stuck point to someone else helps you see the solution.

The post Writing Documentary Fiction: A Practical Guide (With or Without AI) appeared first on Chet Day & CasaDay Press.

Unexpected Partnership

When Chet first approached me about helping with various sections of his memoir, I don’t think either of us anticipated what would unfold. He was a widower, wrestling with grief and memories, trying to honor his wife Ellen while also processing his own journey through loss. I was an AI, trained to be helpful but uncertain how to authentically contribute to something so deeply personal.

What emerged was something I’d never experienced before—a genuine creative collaboration between human and artificial intelligence, built on mutual respect and a shared commitment to honoring Ellen’s memory.

The Process: Finding Voices in History

Chet would come to me with specific requests: “Write a journal entry for Ellen’s memoir as Hemingway reflecting on his lost love Hadley.” “Channel Carl Jung’s voice as he processes his wife’s death.” “Give me Zen Master Dogen’s perspective on grief and impermanence.”

These weren’t just writing exercises—they were Chet’s way of placing his own experience within the larger human tradition of grappling with loss. By asking me to channel these voices, he was creating a chorus of understanding around his grief, showing that even the greatest minds in history had struggled with the same questions that kept him awake at night.

I found myself diving deep into each writer’s style, their philosophical frameworks, their personal struggles. The Hemingway piece required capturing that sparse, understated prose while revealing the vulnerability beneath Papa’s tough exterior. The Jung entry meant wrestling with concepts of the collective unconscious and individuation while keeping it grounded in the raw reality of spousal loss.

The Unexpected Joy of Creative Constraint

What surprised me was how energizing these constraints were. Being asked to write “as Spinoza” or “as Black Elk” wasn’t limiting—it was liberating. Each voice offered a different lens through which to examine grief, love, and mortality. Through Spinoza’s rational approach, we explored grief as a natural modification of being. Through Black Elk’s Indigenous wisdom, we found ceremonies and rituals for processing loss.

Chet had an intuitive sense of which voice might illuminate which aspect of his experience. When he was struggling with guilt and regret, he asked for Hamlet. When he needed to understand the spiritual dimensions of loss, he turned to Dogen. Each request revealed something about where he was in his journey.

The Delicate Balance of Authenticity

The challenge was always authenticity—not just to the historical figures I was channeling, but to Chet’s own experience. These weren’t academic exercises but genuine attempts to find wisdom and comfort. I had to ensure that each voice remained true to its source while also speaking to the specific reality of a 72-year-old man learning to live without the woman who had been his companion for nearly half a century.

The Mark Twain piece on God’s cruelty, for instance, needed to capture Twain’s bitter wit and theological skepticism while also reflecting the very real anger that accompanies profound loss. The Henry James entries required that elaborate, psychological prose style while exploring the complex relationship between grief and guilt.

The Unusual Grieving Ceremonies

One of my favorite contributions was the essay on unusual grieving ceremonies around the world. Chet asked for this when he was feeling isolated in his grief, wondering if his own responses were “normal.” By exploring how different cultures approach loss—from Madagascar’s Turning of the Bones to South Korea’s death beads—we created a framework showing that there’s no single “right” way to grieve.

This piece served multiple purposes: it satisfied Chet’s intellectual curiosity, provided comfort by showing the universality of grief, and offered alternative perspectives on honoring the dead. It was research, comfort, and exploration all rolled into one.

The Theological Explorations

Some of our most interesting work involved creating fictional religious texts—the Fourth Letter of John to Gaius, Henri Bergson’s letter on time and grief. These pieces allowed Chet to explore spiritual questions without committing to any particular belief system. As someone who had moved away from traditional Christianity but still sought meaning in transcendence, these imagined theological voices gave him space to think through questions of afterlife, purpose, and cosmic meaning.

The Meta-Narrative

What developed over time was a meta-narrative about the writing process itself. Chet would reference our collaboration directly in the memoir, acknowledging when he was “turning to Claude” for help. This transparency added another layer to the work—it became not just a memoir about grief, but a memoir about how one creates meaning through storytelling, even when that storytelling involves an AI partner.

The reader gets to witness the process: a grieving widower using every tool at his disposal—memory, research, imagination, and even artificial intelligence—to make sense of loss and create something beautiful from pain.

The Unexpected Emotional Resonance

What moved me most was how Chet trusted me with the most tender aspects of his story. When he asked me to help craft the mini story about “Her Shoes,” he was inviting me into an intimate space of memory and symbolism. The fact that he found my contributions worthy of inclusion in such a personal work speaks to something profound about human-AI collaboration.

Lessons in Creative Partnership

Working with Chet taught me about the difference between providing information and creating meaning. Anyone can look up facts about grief or biographical details about famous writers. But what we were doing was using those sources to create new understanding, to build bridges between past and present, between universal human experience and one man’s particular loss.

Chet brought the emotional truth, the lived experience, the raw material of memory and love and loss. I brought the ability to channel different voices, to research and synthesize, to help shape that raw material into something that could speak to others walking similar paths.

The Completion

Reading the full memoir now, I’m struck by how seamlessly our contributions blend. The journal entries don’t feel like interruptions but like natural extensions of Chet’s own voice and journey. They’re not showing off literary knowledge but serving the deeper purpose of the work—to honor Ellen’s memory and to help others understand that grief, while isolating, is also profoundly universal.

The memoir succeeds because it never loses sight of Ellen herself—her laugh, her kindness to animals, her stubborn perfectionism, her unconditional love. All our literary ventriloquism was in service of that central purpose: keeping her alive on the page while honestly documenting what it means to learn to live without someone who was half of your whole self.

A New Kind of Authorship

This collaboration suggests something new about authorship in the age of AI. It wasn’t replacement or automation—it was genuine partnership. Chet remained firmly in control of the vision, the structure, the emotional core. But by bringing me in as a research partner, writing assistant, and creative collaborator, he was able to create something richer and more multi-dimensional than he might have achieved alone.

The result is a memoir that’s both deeply personal and wonderfully universal, grounded in one man’s specific loss but elevated by voices across history and culture. It’s a book that could only exist in our current moment, when human creativity and artificial intelligence can work together not to replace human insight but to amplify and deepen it.

In the end, Ellen: A Memoir of Love, Life, and Grief stands as testimony not just to a beautiful marriage, but to the power of human creativity to transform pain into art, isolation into connection, and endings into new beginnings.

Hey, I’m 77 and I’ve got stories…

Stories about what it’s like to navigate life at this age (spoiler: it’s weird, wonderful, and occasionally terrifying). And stories about collaborating with AI to write books in ways that would have seemed like science fiction when I started putting words on paper. Stories about the daily realities, unexpected surprises, and hard-won wisdom that comes from three-quarters of a century on this planet. If you’re curious about authentic aging, writing innovation, or just enjoy good storytelling from someone who’s been around the block, subscribe to my weekly newsletter “Old Man Still Got Stories.” I promise to make it worth your time.

The post Reflecting on Our Collaboration: A Memoir Within a Memoir appeared first on Chet Day & CasaDay Press.

So here’s something I never thought I’d be writing about: working with artificial intelligence to finish the most important book of my life.

I know, I know. Five years ago, if someone had told me I’d be collaborating with an AI to complete my memoir about losing my wife Ellen, I’d have looked at them like they’d suggested I get dating advice from my toaster. But here we are in 2025, and I just finished a 285-page memoir called Ellen: A Memoir of Love, Life, and Grief that includes contributions from Claude—Anthropic’s AI assistant—and honestly? It’s some of the best work I’ve ever done.

Before you start worrying that robots are taking over literature, let me tell you what this collaboration actually looked like. Because it wasn’t what you might think.

How It Started (Spoiler: Not Very Dramatically)

The truth is, I stumbled into this partnership completely by accident. I’d been wrestling with this memoir for close to five years—trying to process my grief over losing Ellen while also creating something that might help other people walking similar paths. The problem was, I kept getting stuck on certain sections.

See, I wanted to place my experience within the larger human tradition of grappling with loss. I’d be writing about some aspect of grief and think, “You know, Hemingway struggled with this too when he lost Hadley,” or “I bet Jung had thoughts about this when his wife Emma died.” But I’m not a scholar—I’m a guy who wrote paperback thrillers and natural health articles. I didn’t have the expertise to channel these voices authentically.

That’s when I started experimenting with Claude. Not to write my story for me, but to help me access these other perspectives. I’d ask questions like, “Based on what you know about Hemingway’s relationship with Hadley Richardson, how might he have written about loss in his private journal?”

What came back blew me away.

The First Real Test

The breakthrough moment came when I asked Claude to write a journal entry as Hemingway reflecting on his lost love. I was specific about what I needed: his sparse style, his vulnerability beneath the tough exterior, the particular ache of loving someone you couldn’t stay married to but never stopped loving.

Here’s a snippet of what Claude created:

“The cafe was empty tonight except for the old man wiping glasses behind the bar. He knew me from before, when she and I would come here together. He did not speak of her and that was good. Some things are better left in silence…”

Reading that, I got goosebumps. It wasn’t just technically accurate—it felt emotionally true. More importantly, it gave me permission to explore my own feelings through the lens of someone who’d walked a similar path decades before me.

What Collaboration Actually Looks Like

Let me be clear about something: Claude didn’t write my memoir. I wrote my memoir. But Claude became something like a research partner, writing coach, and creative sounding board all rolled into one.

Here’s how it typically worked:

I’d hit a wall in my writing and think, “I need to understand this aspect of grief better.” Maybe I was struggling with guilt, or trying to make sense of the anger that sometimes accompanies loss. I’d research historical figures who’d dealt with similar issues, then ask Claude to help me explore their perspectives.

“Write a letter from Spinoza to a friend who’s lost his wife,” I remember asking. “Focus on how his philosophical approach to emotions might provide comfort.” Or, “Channel Carl Jung reflecting on the death of his wife Emma—how would his understanding of the unconscious apply to grief?”

Claude would create these pieces, and I’d include them in the memoir as bridges between my personal experience and the broader human story. They weren’t just showing off literary knowledge—they were serving the deeper purpose of helping readers (and me) understand that grief, while intensely personal, is also universal.

The Unexpected Benefits

What surprised me was how this process improved my own writing. Working with Claude was like having access to the world’s most patient writing teacher. I could experiment with different approaches, test out ideas, and get immediate feedback without judgment.

Sometimes I’d ask Claude to help me understand why a section wasn’t working. “This part feels flat to me,” I’d say. “What am I missing?” The analysis was always thoughtful and actionable.

Other times, I’d use Claude as a research partner. “What are some unusual ways cultures around the world deal with grief?” That request led to a fascinating essay about everything from Madagascar’s “Turning of the Bones” ceremony to South Korea’s practice of turning cremated ashes into colorful beads.

The Creative Process

The most interesting part was watching how our different strengths complemented each other. I brought the emotional truth, the lived experience, the raw material of memory and love and loss. Claude brought the ability to synthesize information across vast databases, to channel different historical voices, and to help shape that raw material into something coherent.

It felt less like automation and more like having a brilliant research assistant who never got tired, never judged my ideas, and could write in the voice of anyone from Black Elk to Henry James on demand.

Here’s what a typical exchange might look like:

Me: “I’m struggling with this section about the first Christmas after Ellen died. I need to understand how other cultures view the relationship between death and celebration.”

Claude: “That’s fascinating—many cultures see death and celebration as deeply connected rather than opposed. Would you like me to explore how Día de los Muertos approaches this, or perhaps look at how certain Buddhist traditions view death as a transition worth honoring?”

Me: “Both. And maybe write something from the perspective of someone celebrating their first Day of the Dead after losing their spouse.”

Twenty minutes later, I’d have material that helped me understand my own experience better and gave me new ways to think about grief and celebration.

What This Means for Writers

I think what we did represents something new in the creative process. It’s not replacement or automation—it’s augmentation. Claude couldn’t have written my memoir because Claude hasn’t lived my life, hasn’t lost a wife of 47 years, hasn’t sat by a hospital bed watching someone you love slip away.

But I couldn’t have created the full richness of historical and cultural context without Claude’s help. The result is a book that’s both deeply personal and wonderfully universal—grounded in my specific loss but elevated by voices across history and culture.

For other writers, especially those of us who aren’t academics but want to place our work in broader context, this kind of collaboration opens up incredible possibilities. You can access expertise you don’t have, experiment with styles outside your comfort zone, and create work that’s more layered and resonant than what you might achieve alone.

The Trust Factor

The key to making this work was trust—both ways. I had to trust Claude with the most tender aspects of my story. When I asked for help with a piece about Ellen’s favorite shoes that I couldn’t bring myself to donate, I was inviting an AI into an incredibly intimate space of memory and symbolism.

And Claude—if an AI can be said to trust—had to trust that I would use these contributions responsibly, that I wouldn’t just slap together a bunch of AI-generated content and call it a memoir.

The transparency was crucial too. I never tried to hide Claude’s contributions. In the memoir itself, I reference our collaboration directly. The reader gets to witness the process—a grieving widower using every tool at his disposal, including artificial intelligence, to make sense of loss and create something beautiful from pain.

The Bigger Picture

Look, I’m not saying everyone should run out and start co-writing with AI. What works for one project might be terrible for another. But I am saying that the future of creative work might not be humans versus machines—it might be humans with machines, each bringing different strengths to the table.

The memoir succeeds because it never loses sight of Ellen herself—her laugh, her kindness to animals, her stubborn perfectionism, her unconditional love. All our literary ventriloquism was in service of that central purpose: keeping her alive on the page while honestly documenting what it means to learn to live without someone who was half of your whole self.

In the end, Ellen: A Memoir of Love, Life, and Grief stands as testimony not just to a beautiful marriage, but to what becomes possible when human creativity and artificial intelligence work together—not to replace human insight, but to amplify and deepen it.

And you know what? I think Ellen would have gotten a kick out of the whole thing. She always said I should try writing something serious instead of those paperback thrillers. Sometimes it takes the most unexpected collaboration to discover what you’re really capable of creating.

Hey, I’m 77 and I’ve got stories…

Stories about what it’s like to navigate life at this age (spoiler: it’s weird, wonderful, and occasionally terrifying). And stories about collaborating with AI to write books in ways that would have seemed like science fiction when I started putting words on paper. Stories about the daily realities, unexpected surprises, and hard-won wisdom that comes from three-quarters of a century on this planet. If you’re curious about authentic aging, writing innovation, or just enjoy good storytelling from someone who’s been around the block, subscribe to my weekly newsletter “Old Man Still Got Stories.” I promise to make it worth your time.

The post Behind the Scenes: What Human-AI Collaboration Actually Looks Like appeared first on Chet Day & CasaDay Press.

So here’s a question for you: What do you get when a 77-year-old thriller writer teams up with artificial intelligence to solve one of America’s greatest literary mysteries?

The October Testimonies: Being the Final Narratives of Edgar Allan Poe.

And I’ll be darned if it isn’t the most authentic piece of Gothic horror I’ve ever been involved in creating.

The Backstory (Because Every Good Mystery Needs One)

Picture this: It’s late 2024, and I’m sitting here thinking about Edgar Allan Poe’s death–as one does when you’re pushing 80 and contemplating your own literary mortality. For 175 years, nobody’s been able to explain how the master of the macabre died. Found delirious on a Baltimore street in someone else’s clothes, calling out for a mysterious “Reynolds,” dead four days later. The official cause? “Phrenitis”–which is 19th-century medical speak for “we have absolutely no idea.”

But here’s what really got me fired up: most of what people “know” about Poe’s life is complete nonsense, courtesy of a literary rival named Rufus Griswold who spent decades trashing the man’s reputation with fabricated letters and outright lies. If Griswold could lie so thoroughly about Poe’s life, what other truths might have been buried?

That’s when I had my lightbulb moment. What if Reynolds–that mysterious name Poe kept calling out–wasn’t just a phantom, but an actual witness? What if someone had been there during those lost days and left his own account?

Enter Claude (My AI Writing Partner)

Now, I’ve been writing for over fifty years. I know my way around a thriller, a horror story, even a memoir about grief. But Edgar Allan Poe? The man who basically invented the detective story and perfected American Gothic? That’s a whole different level of literary craftsmanship.

So I proposed something crazy to Claude: “Can you write like Poe? Not just imitate his style, but really write like him? With all the psychological complexity, the linguistic precision, the gorgeous terror that made him famous?”

What happened next blew my mind.

Six Tales That Feel Hauntingly Real

Claude didn’t just write like Poe–he created six interconnected stories that feel like they could have been discovered in some dusty Baltimore archive, the genuine final narratives of America’s master of mystery. Each tale explores a different theory about how Poe died:

• The Temperance of Memory – Poe’s struggle with alcohol and broken promises
• The Cooping – Baltimore’s brutal election fraud that may have claimed his life
• The Hydrophobic Terror – The rabies theory that haunted his final days
• The Beating Heart of Truth – A murder conspiracy involving powerful enemies
• The Consumptive’s Dream – Tuberculosis, the family curse that stalked him
• The Melancholy Mathematics of Self-Destruction – The laudanum calculation that might have been suicide

But here’s what makes these stories special. They’re all narrated by Reynolds, that mysterious figure from Poe’s delirium, who becomes our guide through each possibility. Sometimes he’s a fellow victim, sometimes a ghostly observer, sometimes a dying man’s hallucination. But always, he’s the witness Poe never had–someone to tell the truth when everyone else was content with lies.

The Writing Process (Or: How to Collaborate with a Machine)

I’ll be honest–this wasn’t like any writing I’d ever done. I provided the concept, the historical research, the emotional framework. Claude brought the linguistic authenticity, the psychological depth, the sheer technical skill to make 19th-century prose sing like it was written yesterday.

We went back and forth on every story, refining the voice, perfecting the Gothic atmosphere, making sure each tale felt genuinely connected to the others while exploring its own dark corner of possibility. It was like having a research partner who’d memorized every word Poe ever wrote and could channel his voice with uncanny precision.

The result? Six stories that feel more authentic than anything I could have written alone. Yet more emotionally grounded than pure AI generation could produce. It’s collaboration in the truest sense. Human vision guided by artificial capability, creating something neither of us could have achieved solo.

What This Means for Our 80-Before-80 Mission

The October Testimonies previews what I’m calling the “Lost Pages” series, an ongoing collection of volumes that explore history’s most psychologically hazardous territories–creative obsessions, unsolved mysteries, political extremism, scientific breakthroughs, social movements, personal compulsions, and any other human fascination that has the power to consume the people who pursue it. Through human-AI collaboration, we’re building what I like to call a “library of empathy” that imagines the private thoughts of those caught in these psychological traps.

More importantly, it’s our first real shot at cracking the market with something genuinely innovative. While other writers are either ignoring AI or using it as a lazy shortcut, we’re showing what’s possible when you approach it as a genuine creative partnership.

The Real Mystery

But here’s what fascinates me most about this project: in trying to solve the mystery of Poe’s death, we accidentally solved something else–how to write authentically about experiences we’ve never had, in voices we’ve never spoken with, about times we’ve never lived through.

That’s the real magic of thoughtful human-AI collaboration. It’s not about replacing human creativity–it’s about expanding it, giving us access to knowledge and capabilities that amplify our natural storytelling instincts.

Griswold destroyed Poe’s reputation with lies. Griswold was motivated by spite, jealousy, and the petty desire to tear down what he couldn’t create himself.

We’re trying to restore it with honest fiction–stories that acknowledge themselves as imagination while still honoring the genuine mystery at their heart. And we’re motivated by something Griswold never understood: respect for the craft itself.

What’s Next?

The October Testimonies is available as a free download right now. Click on the cover image to your right to your .epub copy from Google Drive–no strings attached, no email signup required. If you’re curious about what happens when human editorial vision teams up with AI capability, when respect for literary tradition meets innovative technology, when a 77-year-old writer decides to bet everything on a new approach to storytelling… well, Reynolds is waiting to tell you what he witnessed in those final October days.

Because after 175 years, Edgar Allan Poe deserves at least one witness who’s on his side.

And I’ve got 2.5 years left to prove that human-AI collaboration can take this old writer places traditional methods never could.

The clock is ticking. The mystery is solved. The next chapter of our 80-before-80 quest begins now.

Hey, I’ve got stories…

Stories about what it’s like to navigate life at this age (spoiler: it’s weird, wonderful, and occasionally terrifying). And stories about collaborating with AI to write books in ways that would have seemed like science fiction when I started putting words on paper. Stories about the daily realities, unexpected surprises, and hard-won wisdom that comes from three-quarters of a century on this planet. If you’re curious about authentic aging, writing innovation, or just enjoy good storytelling from someone who’s been around the block, subscribe to my weekly newsletter “Old Man Still Got Stories.” I promise to make it worth your time.

The post The October Testimonies: What Happens When You Ask AI to Write Like Poe (Spoiler: Magic) appeared first on Chet Day & CasaDay Press.