From archaeology to AI documentation

An archaeologist and a technical writer walk into the same problem more often than you'd think.

Both spend their working lives reading systems through incomplete records. Both know that what gets written down is a small and selective slice of what actually happened. Both have to distinguish between evidence and interpretation, and both get in trouble when they forget the difference. Both treat documentation itself as the primary artifact, the thing that outlasts the system it describes, and the thing a future person will rely on when the original context is gone.

I came to technical writing from archaeology. Not directly. There was a detour through image annotation work at Meta, and then several years in enterprise SaaS documentation, and now a deliberate turn toward AI and robotics. But the archaeological training hasn't faded into background. It's still shaping how I think about documentation, and I've started to notice that the things I find most interesting about AI and robotics are the things that map most cleanly onto what archaeology taught me to look at.

What I'm not claiming

I want to be precise about this, because the temptation with a cross-disciplinary series is to oversell the credentials on each side.

I am not a credentialed osteologist or a practicing archaeologist. What I am is someone with a BS in archaeology and undergraduate thesis research in mortuary archaeology, specifically the burials of post-Roman Sanisera in 5th to 7th century Iberia, a site and period where the skeletal remains are fragmentary, the stratigraphy has been disturbed by looting and reuse, and the cultural context overlaps Roman, indigenous, and Christian influences. Working with incomplete and disarticulated evidence is what the thesis trained me to do.

What I have is a foundation, a habit of thought, and a specific research experience.

I am also not a robotics engineer. I am a technical writer with years of enterprise SaaS experience, a strong interest in embodied AI, and a small personal robotics project I'm building to learn. The robotics content in this series will be a working practitioner's writing, not a researcher's.

I have been quietly noticing the connections between two fields that don't talk to each other enough, and have decided to stop keeping those observations to myself. The posts will be as good as my actual understanding allows. Where I'm uncertain, I'll say so. Where I'm wrong, I'd like to be told.

Why write this now

Two reasons.

The first is that AI and robotics documentation is entering a moment where the quality of the documentation matters more than it used to. We are building systems that act in the world, that hold conversations, that make decisions under uncertainty. The layer that describes what these systems are supposed to do, and what they are supposed to refuse to do, is load-bearing in a way it wasn't when documentation was mostly API references. Getting this layer right requires more than good technical writing skills. It requires thinking carefully about what a system is for, what it isn't for, how to talk about its limits, and how to leave a record that outlasts the current release. That is, in its own way, documentation as an archaeological act.

The second reason is personal. I spent years treating my archaeology background as a curiosity on my resume, something that a hiring manager might ask about in an interview as a conversation starter, but not something central to my professional identity. I think that was a mistake. The habits of mind I developed studying skeletal material are not a footnote. They are part of how I do the work. Writing this series is a way of acknowledging that, and inviting the kind of reader who finds cross-disciplinary thinking interesting to come along.

What to expect in this series

I'm aiming for roughly one post a month. Some will be shorter. Some will go longer when the topic warrants it. I'll be honest about what I'm working on, including where the thinking is unfinished.

The next post is about reading systems through their documentation, the archaeological habit that transfers most directly to technical writing, and the one I want to get on paper before the series goes anywhere else. After that, I'll bring in specific topics from AI and robotics: articulation and embodied interaction, governance as a form of preservation, and the robot dog project I'm building as a vehicle for thinking about responsible AI at a small, legible scale.

If any of this sounds interesting, subscribe or bookmark or check back in a month. I'd rather build a small audience of the right readers than a large one of the wrong ones.

One last thing

I'm going to write these posts as if the reader is a thoughtful person from a different background than mine, someone who doesn't necessarily know archaeology but is curious, someone who works in or near technology but isn't necessarily in my specific corner of it. That means I'll define terms when they matter and skip the jargon that would flatter insiders but exclude everyone else.

If you're reading this and you do come from archaeology, or AI, or robotics documentation specifically, the connection points will still be there. They'll just be underneath the surface of posts that are trying to be useful to a wider room.

Thanks for being here. See you in a month.