Walking the dog with Claude — Matt Godbolt’s blog
Walking the dog with Claude
Written with LLM assistance.<br>The interview format is genuine; the prose is lightly tidied from voice notes.
I had lunch with a pal yesterday, and we got onto the subject of why<br>so much technical material is either accurate-but-impenetrable or<br>polished-but-slightly-wrong. It’s a gap I think about a lot, partly<br>because I make videos that try to land in the middle of it and don’t<br>always succeed. The conversation stuck with me on the drive home.
This morning, whilst walking the dog, I tried something a bit silly: I<br>asked Claude (dictating to the mobile app1) to interview me about it.<br>Voice-to-text on one end, an LLM<br>playing journalist on the other, and a black lab setting the pace. What<br>follows is that conversation, lightly tidied – paraphrased in places,<br>verbatim in others. The LLM is the interviewer; I’m the subject.
The words are mine, the questions aren’t.
Claude: Let’s start wide. Your friend’s comment was about the gap<br>between people who know things and people who can explain them. Where<br>do you see that gap most clearly?
Matt: Almost everywhere, honestly. The people with the information<br>may not be great presenters, and the great presenters don’t necessarily<br>have the information. You end up with two failure modes: dry, accurate<br>material that nobody wants to watch, or polished material that gets the<br>details slightly wrong in ways that make experts twitch. Both are<br>common; the overlap is rarer than I’d like.
Claude: Who do you think bridges it well?
Matt: Grady Hillhouse on Practical Engineering – he actually is a<br>civil engineer, and it shows. Devin Stone, LegalEagle, is a working<br>lawyer. And the Brady Haran shows – Numberphile, Computerphile – solve<br>it from the other end: Brady and Sean Riley are well-prepared “straight<br>man” interviewers, and the experts they bring in are carefully chosen.<br>The format itself is the trick. I’m biased on Computerphile, mind, since<br>I’ve done a few.
Claude: Let’s talk craft. You’ve mentioned the Usborne robots<br>before. What are they, and why do they still matter to you?
Matt: There were these late-70s / early-80s Usborne books on how<br>computers worked, and they used little cartoon robots to show<br>information moving around inside the machine. As a kid that just<br>clicked for me. The thing I find quietly remarkable is that I still<br>use it. When I’m reasoning about out-of-order execution, register<br>renaming, modern microarchitecture – somewhere underneath, there are<br>still little robots shuffling things between pigeon-holes. I’m not<br>consciously picturing them anymore, but the shape of the model is the<br>same.
Claude: That sounds like a beginner’s model that you’ve never quite<br>outgrown.
Matt: Maybe you’re right? The thing I<br>keep coming back to is that what’s easy for me to reason about as an<br>expert2 is often exactly what’s easy for a beginner to grasp. The<br>mental model and the explanation aren’t two things – they’re the same<br>object. I think people assume good explanations are simplifications of<br>the expert’s real model. Sometimes they’re just the expert’s real<br>model.
Claude: You’ve talked about a “tell my mum” test. Where does that<br>come from?
Matt: When I was a teenager with a BBC Micro, if I figured<br>something out I’d run downstairs and tell my mum. She’d listen and<br>either follow or politely pretend to. Somewhere along the way that<br>became my actual yardstick: could I explain this to her? Simple enough<br>to follow, but not dumbed down. True. And underwritten by the fact<br>that I actually understand the thing properly, even if I don’t expose<br>all of it. If I can’t clear that bar, I’m not ready.
Claude: And the prep that goes with it?
Matt: More than people probably realise. For the early<br>Computerphiles I’d practise on paper, out loud, two or three times,<br>even though the finished video looks off-the-cuff. The<br>off-the-cuffness is the prep working.
Claude: You skipped a Computerphile recording last week. What<br>happened?
Matt: Yeah, that’s been nagging at me. The topic was how computers<br>tell the time – it’s fallout from my C++Now keynote on benchmarking5.<br>I wanted to talk about rdtsc and rdtscp on x86, how raw counter<br>ticks get turned into wall-clock time, and the out-of-order execution<br>wrinkles that show up when you read the clock. Good material. I just<br>couldn’t find a visual hook. Computerphile’s whole stick is hand-drawn,<br>handwritten stuff on lined paper, and I hadn’t worked out how to draw<br>this one. I’m now mulling writing a tiny x86 assembly snippet<br>long-hand and using that to show out-of-order behaviour. But the honest<br>reason I bailed is that I hadn’t cleared the mum test yet. I hadn’t<br>reduced it cleanly enough.
Claude: Why do you do these at all?...