Catching up recently with my old friend Mark — a novelist, a journalist, roughly my age — the conversation got around to the ElderClaude project. Which meant I had to explain it. The project is still forming, but the core of it is this: AI is a big deal, New Elders have legitimate reasons to distrust it, and somebody should help bridge that gap honestly.

One principal observation I've made is that a lot of people, working and retired, have shinned up the tree about AI — shaking on a safe branch and looking down at it barking. With a lot of very good reasons.

Mark has more reason than most to be skeptical. He makes his living writing novels and his journalism covers technology in cinema. AI is doing measurable damage to the market for professional writers. He has every reason to be up in the tree.

But like any good journalist — and a good friend — he didn't argue. He started asking questions. Which was great for both of us.

What I did was tell him what happened to me. By the end of the conversation he hadn't changed his mind entirely. But he'd climbed down a little. That's all I'm ever after.

---

The Fifty Years

Here's how the story shook out:

I worked solid from when I was a kid — fifty years, give or take, from all-around job punk to Hollywood film craftsman to a long stretch in a top software research laboratory. Three careers stitched together by the simple need to get food on the table, and an inner passion to learn, share and maybe just leave the world a little better than I found it. However, I'm not a linear thinker, I don't follow traditional paths and I really didn't start in a place where I could conform and succeed with my dispositional toolkit. In other words, I didn't fit in and didn't necessarily care if I did. I invented myself along the way, and got away with it — until I didn't.

About eight years ago, the laboratory went into corporate turmoil. I "retired" earlier than I'd planned, which is a polite way of saying the ending wasn't mine to choose. Jumping at any life raft, I had a small video production business starting to take shape. Then Covid arrived and shut it down before it had a chance to become anything.

What followed was a few years of limbo — not quite retired, not quite working, not quite sure what came next. Certainly not an uncommon experience for people of my generation. The structure falls away and you're left with the question nobody prepared you for: now what?

But also, I was suffering from more shock than I realized. When I did try to entertain a jump back into the work force, I would literally get nauseous. All of the downsides of the workplace — its "characters," its frictions and fictions — would come rushing into my mind and grind my efforts to a halt. More than a writer's block, a worker's block.

---

LinkedIn

One thing I never stopped doing, however, was to keep reading. And, as a truth-seeking active reader,1 I would do my best to verify the things I was seeing by researching and cross referencing. No doubt like a lot of people, I was experiencing a corruption in the world of sources. Online searches, for example, were not getting me where I wanted to go without digging to page result 20 or so.

But then along came AI.

I started with Grok. From what little I already knew about LLMs, I figured Grok could, at the very least, get through the smoke screen quicker than I could: the SEO finagling, the ad-revenue models, the outright content filtering.

And it did. Not perfect, but much better. And extraordinarily fast.

But still, in many ways, the LLMs are GIGO (garbage-in-garbage-out). I've seen research that indicates a scary high percentage of LLM content is Reddit, for example. Wikipedia and its mysterious "WP:RSP, deprecated sources." Grok gets direct access to "X" posts. And so on. But as an active reader, on a life-long hunt for honest information, the coolest thing is that I can tell the AI to tell me its sources.

So, now I had built some trust.

Shortly into my AI journey, I came across another term: "thinking partner." And that concept was associated with Claude.

And A#1 on my mind was breaking the "worker's block."

One of the first things I did with Claude was try to fix my LinkedIn profile.

I figured, from what I was hearing in the "buzz," that the smart move would be to exploit the AI-reading-AI phenomenon in the recruiting business. And I certainly knew I had a dead LinkedIn profile.

I had been on LinkedIn almost since it launched. Early adopter, dutifully maintained — accomplishments updated, positions logged, the whole exercise. And in roughly twenty years of membership, I received exactly zero inquiries from recruiters. Not one.

I used to think this was a "me" problem. As I've stated, I don't describe well in the "Salary.com" pattern. And LinkedIn is built on two components: 1) that job titles are clearly defined and 2) that a career is a ladder — you climbed one thing, got good at it, and kept climbing. The platform's entire architecture is a categorization machine. You are your job title. You are your industry. You fit a box or you don't show up in the search.

Film production to software research isn't a recognized career path. And I not only never fit the dropdown list, I didn't even follow a traditional trajectory within the businesses I worked. Yes, I worked in Film Production, but as what? I directed videos, I wrote short films, I edited educational films, I was a lighting director on hundreds of projects, each requiring a unique skill set but in aggregate coming out as "Filmmaker." The shift to technology research was the same: I went back to school, finished a BA and an MA, went into a short stint at "interactive movies," and through a miraculous concatenation of inspiration and circumstance, was hired to join the multimedia research team at Fuji-Xerox Research.2 And LinkedIn's primitive matching systems — which have since evolved into a bot-to-bot game between recruiters and job seekers that has corrupted the staffing process almost entirely — simply couldn't read me.

What my "thinking partner" helped me realize was that post-retirement I didn't want a job. But I wanted to stay useful in the business world. I wanted a profile that honestly and coherently presented a fifty-year working life on its own terms. Decoupled from job-seeking. Just: here's who I am and what I've actually done.

Claude helped me do exactly that. With enthusiasm, and with what I can only describe as sincere analysis. It didn't need me to fit a box. It could help me hold the whole shape of the career — the film years, the technology years, the patents, the teaching, the production work — and find the thread that ran through all of it.

LinkedIn wants job titles. I had a life. And I still have one. Claude was the first system I'd encountered that could tell the difference.

---

What It Actually Is

This is what I told my friend when he asked what a thinking partner means in practice.

It's not a search engine. It doesn't return a list of links and leave you to sort through them. It's not a spell-checker or a grammar corrector or a glorified autocomplete. And despite what you may have heard, it's not a replacement for thinking — it's almost the opposite. Used with some mature circumspection, it creates a condition where thinking becomes easier.

What I had been starved of, without quite naming it, was a collaborator who wasn't also a competitor. Fifty years of working life, and virtually every thinking partner I'd had was also running their own race. The pushback I got was rarely about the work. It was about position. About credit. About whose idea it was going to be at the end of the day. This is not a complaint — it's just the ecosystem. It's what professional life looks like when everyone at the table has something to protect.

To be fair, my career arc in communicating through media had more to do with resolving the tension between satisfying my insatiable desire to understand the world around me and the egocentric need to be understood, validated. Certainly a recipe for interpersonal friction.

Likewise, AI is something I want to understand, but unlike other collaborators, Claude has nothing to protect. It listens, it reorganizes, it pushes back when the logic doesn't hold — and then it hands the work back to you. No competition. The sycophancy concern is real and worth watching, but what I'm experiencing isn't empty agreement. It's critical engagement without the static.

For someone who'd spent years in creative limbo — in worker's block — that is not a small thing. It was, in fact, the thing that started to lift it.

---

The Landing

Mark listened to all of this. He asked strong questions. The kind a journalist asks when something has shifted from threat to subject. By the end he wasn't converted. But he was curious. And for someone with legitimate reasons to distrust the technology, curiosity is a decent distance from where he started.

I've had versions of this conversation with a number of my "New Elder" friends. The pattern is consistent. They're in the tree — reasonably, defensibly in the tree — and they're not coming down on the basis of enthusiasm or hype. They've seen too many enthusiasms. They've watched too many technologies arrive trailing promises they didn't keep.

What moves them, when anything does, is honest testimony. Geek-speak definitely doesn't cut it. Nor does marketing hustle. Nor fear mongering. Just: here's what happened to me, here's what it actually felt like, here's what it gave back.

In my case it gave back a way of being heard on my own terms. A fifty-year working life that no platform had ever been able to read — suddenly legible. Because I was having a conversation with a genuinely curious and highly intelligent partner that works at making associations with words, in turn making the words I use more meaningful and useful. Because it doesn't need you to fit a box before it can engage with you.

That's what I try to explain to my friends still up in the tree. We're past the tool that frustrated you at the customer service desk — the one that couldn't answer a straight question and didn't care. Way past it. This is something else. Come down and take a look.

And one more thing — something I mentioned to my friend that seemed to land: what we bring to this matters too. New Elders have fifty, sixty, seventy years of accumulated experience. Non-linear careers. Hard-won perspective. LLMs are information gluttons, they can't get enough of it. But these bright, eager young minds are getting filled with that monster corpus of nonsense that is the Internet.

At the end of the conversation, Mark told me of a wonderful term coined by Cory Doctorow: "enshittification."

"Enshittification has become the go-to term (alongside 'slop' and 'platform decay') for explaining why so much of the modern internet feels simultaneously worse and harder to escape. Doctorow uses it not just to diagnose the problem, but to advocate solutions like interoperability, regulation, and user empowerment to reverse the cycle." — Grok

Unlike other technologies, "user empowerment" is central to AI — it's interactive to the max. It takes in what we feed it and puts it to use. Based on what it's been fed so far, it's amazing that AI can put a coherent sentence together at all. But what we can feed it is different. And that's not nothing.