BusyDoggie Fieldnotes

BusyDoggie is something I do in my spare time because I enjoy investigating things, especially things about dogs, and I think sharing useful information and knowledge with other people is a worthwhile thing to do.

I'm not a dog professional and nobody pays me to do this. My background is in research, so when something catches my attention my instinct is usually to start asking questions, looking for evidence and trying to work out what is actually going on.

Sometimes that eventually becomes a BusyDoggie article, but producing them takes time. Even with AI helping with parts of the research and writing process, a substantial article can take me several days to research, check, think through and write properly. If someone is going to give me their time and attention, I want to respect that.

However, I investigate and ponder far more things than I can ever turn into articles, partly because there are only so many hours in the day and partly because I also need to write extensively in my actual job. 😅

This is a public notebook for some of that thinking: questions, observations, useful distinctions, bits of evidence and connections I want to keep rather than lose simply because I never found the time to turn them into a finished piece of writing.

A surprising amount of this starts while I'm doing something else. Something piques my curiosity and I start talking it through with ChatGPT. I might be sitting at my desk, but I might equally be walking the dogs or on my way somewhere.

I often use voice conversations to process my thoughts: asking questions, following evidence, comparing explanations, testing an interpretation and gradually working out what I think.

Some of these explorations become very long. Useful ideas used to disappear somewhere inside them. I might later find myself discussing one of those ideas with someone and know that I investigated it in considerable detail, while being completely unable to retrieve the details without searching through the equivalent of hundreds of pages of conversation.

This collection is my attempt to fix that.

Isn't AI unreliable?

Well, my working relationship with AI is rather different from typing a question into ChatGPT and treating whatever comes back as the answer.

I'm a researcher by background, I have four master's degrees across different fields, and AI is now part of my professional work. Over the past year I've also been studying for a master's in AI governance, alongside exploring how people can work effectively and responsibly with intelligent technologies. I therefore spend a perhaps unreasonable amount of time thinking about what these systems can do, where they fail, and how to use them without handing over your judgement to the machine.

Over time, I've built dozens of specialised versions of ChatGPT for different kinds of research and reasoning. I give them detailed instructions about how to work, and some have their own libraries of research and reference material. Depending on the task, they might be instructed to trace claims back to sources, look for contradictory evidence, distinguish correlation from causation, challenge my interpretation, preserve uncertainty, or keep published evidence separate from my own observations and hypotheses.

I also use specialised academic research tools such as Consensus, Scite, SciSpace and Anara which connect directly to ChatGPT. This means I can use dedicated research databases and tools to find, interrogate and cross-check published research within the same conversation, rather than relying on the language model itself to tell me what the literature says.

Of course, none of this makes generative AI infallible. That is exactly why it matters how it is used: I check sources, observations remain observations and hypotheses remain hypotheses. Where evidence is uncertain, I try to preserve that uncertainty rather than force a cleaner answer.

When one of these explorations produces something worth keeping, I ask ChatGPT to help identify the small number of ideas that might still be useful later. I choose which ones belong here, and they are saved as individual entries rather than remaining buried inside the original conversation which I then archive. I also save papers, articles and other sources directly.

Ultimately, this digital garden is mostly for me, but I've made it public in case some of it is useful to someone else too.

Take what piques your interest and leave the rest. 😉