A Reader Stopped Me Cold. Here's What He Said About Grief, AI, and the Brain Nobody Understands.

Grief is not a problem to be solved. It's something you go through. That distinction is everything.

A few weeks ago I published a post about AI resurrection.

About the companies building services that let you have conversations with a chatbot trained on the texts, emails, and voicemails of someone who has died. About a man in our community who lost his wife of 44 years and told me quietly, like he was admitting something he wasn't sure he should, that he was thinking about using one.

I said at the end that I was still working out my own answer. That I held my views loosely.

And then a reader responded in the comments.

And I haven't stopped thinking about what he wrote.

What He Said

I'm not going to share his name, that's his to give if he wants to. But I want to quote him directly, because I don't think paraphrasing does it justice.

He wrote, in part:

"They believe they understand the human psyche, when they do not and never have. They also believe that they understand how the brain works, at best a laughable conclusion. The best neurologists in the world will tell you that they don't understand how the brain works, and that nobody does. It remains an insoluble mystery."

And then, at the end of his comment, this:

"Grief remains a state which one must go through to get out of it. Time heals more than anything."

That last line.

I've read it maybe twenty times since he posted it. Because in one quiet sentence, no jargon, no corporate language, no academic hedging, he broke the entire grief tech business model.

Not a regulatory argument. Not a fiduciary standard.

Just a plain human truth that the product is designed to route around.

The Difference Between Ignorance and Malice

One of the things that hit me hardest in his comment was a distinction I hadn't drawn clearly enough in my original post.

He said he doesn't believe AI researchers are cynical or malicious. He believes they are ignorant. Operating from a framework that has had extraordinary success in well-defined problems; chess, protein folding, image recognition, and extrapolating from that success in ways that aren't warranted.

It worked there. So why wouldn't it work on grief?

Because grief isn't a problem.

That's why.

Chess has a defined state space and a measurable outcome. You can optimize for it. But grief is not a problem with an optimal solution. It is an experience. It changes shape as you move through it. It changes you as you move through it. And the change, the disorienting, slow, non-optimizable change, isn't a side effect to be minimized.

It's the point.

Assuming that if you model the outputs of human emotion accurately enough you've understood the thing itself — that's a particular kind of intellectual overconfidence that comes from having solved the wrong problems too well.

What Neuroscience Keeps Not Knowing

His point about the brain stopped me.

Because I think I'd been soft-pedalling it.

The technology press talks about consciousness like it's a release date away. You'll see breathless coverage of AI models that "understand" human emotion, that "detect" mental states, that are "trained on" human psychology as if psychology were a dataset you could just download and process.

Meanwhile, the best neurologists in the world, the actual researchers who have dedicated their careers to studying the brain, will tell you plainly that they don't understand how the brain works. That every discovery opens new and larger questions. That consciousness remains genuinely, stubbornly beyond the explanatory reach of any current model.

The brain is not a black box we're slowly cracking open. It keeps expanding every time you look inside.

And yet an app was built on the premise that this mystery, the specific way a particular human being loved, grieved, communicated, connected, can be reduced to a training set.

I find that less funny than my reader does. Though his laugh is probably healthier.

The Kübler-Ross Mistake, And Why I Should Have Known This

Here's something I have to admit.

I referenced the five stages of grief in my original post the way most people do. As received wisdom. Without really thinking about where it came from.

My reader pointed out what I should have looked up years ago.

Elizabeth Kübler-Ross developed her model working with people who were dying. Not the people who survived them. Her observations were about the psychological experience of facing your own death. Not about bereavement. Not about the people left behind, who go through a completely different process; one that is often uniquely their own and that doesn't map neatly onto five stages in any reliable sequence.

That model got lifted out of its original context and applied wholesale to grief. Embedded into clinical practice. Into therapy. Into cultural shorthand.

Into the framework that grief tech companies now build their products on.

That's not a small error. When a categorical mistake gets embedded into the tools we use to help people at the worst moments of their lives, it causes real harm. Not because anyone intended harm. Because nobody questioned the assumption.

The same kind of unquestioned assumption is now inside the products being sold on subscription.

The Sentence That Breaks the Business Model

Grief remains a state which one must go through to get out of it.

Not around. Not under. Not past. Not managed with a product designed to make the pain feel less urgent.

Through.

I've written before about AI subscription services having a financial incentive for you to stay dependent. That their business model is not aligned with your recovery. That's true of grief tech in a way that should be said plainly in every conversation about this industry.

If grief is something you must go through to get out of it, a product designed to soften the passage, to give you a place to linger, a simulation to return to, a facsimile of the presence you're missing, is not helping you through.

It's helping you stay.

That might be comforting on a Tuesday evening when the house is quiet and the absence of someone is louder than it should be. I'm not dismissing that. Comfort is real, and sometimes it's what a person needs to get through a day.

But a subscription to comfort is a different thing entirely. And an industry built on that subscription has a direct financial interest in the passage never being completed.

What This Community Has Taught Me

I want to say something that I mean genuinely.

I started this conversation thinking I had a reasonably clear view of the issue.

One reader, in one comment, gave me three things I didn't have before; a cleaner way to think about why the technology fails at grief specifically, an honest correction on something I'd been referencing wrong for years, and the most precise sentence I've heard about what grief actually requires.

That's not something I expected to get from a comment section. And it's exactly why I think this community is worth building.

My reader also noted and he was careful to be clear he's not a clinician, that clinical approaches to grief have genuinely improved. That medication for chronic depression is the best it's ever been. That talk therapy has made real strides. That help exists and is better than it has ever been.

That matters. Because the argument isn't that people in grief have nowhere to turn. The argument is that this particular product, dressed up in empathy, packaged in a subscription, is not what it presents itself as.

And the people most likely to reach for it are the ones with the most to lose from a process that keeps them tethered rather than helping them find their way through.

Before You Go

I want to ask you something honestly.

Did you know that Kübler-Ross's five stages were developed studying the dying, not the bereaved? Because I didn't. Not really. I'd cited them the way everyone cites them, as a general model of grief, without ever asking where they came from.

I'm curious how many of us were operating on the same assumption.

And a separate question, if you've experienced real loss; a spouse, a parent, a close friend, does the idea that grief must be gone through ring true to your experience? Or did something else help more?

Drop it in the comments. I'll read every single one.

This conversation started with one post. A reader made it better. I have a feeling the comments are going to do the same thing again.

➡️ Haven't read the original post? [Read: When AI Brings Back the Dead — And Who Really Benefits].

➡️ Want to talk about this with people who take it seriously? [Come find us in the free Tech 4 Grown-Ups community].

➡️ Prefer to listen? [Catch the conversation on the Tech 4 Grown-Ups Podcast].

— Michael Routhier, Founder of Tech 4 Grown-Ups. I run free digital safety seminars for adults 55+ and write about tech threats as they happen. Learn more about me →