james_ross

The key sentence in this post is when the author says that to understand something, you must actively engage with it. Part of me thinks that saying books don’t work is a bit like saying, say, barbells don’t work. Like barbells, books are passive objects, so expecting them to “work” seems unreasonable. But another part of me agrees that barbells can be designed to be more or less compatible with how human beings grasp and lift them to build strength, and so can books. Having re-read recently “How to Take Smart Notes”, which is all about using the Zettelkasten method to make reading a very diligent and thorough process that ends with every insight from a reading source integrated into your own body of explicit thought outside your brain, another part of me thinks books are fine, and it’s us not putting in the effort to integrate what we read int our existing thoughts that’s the issue. I did not see that book, or books like it such as Novak’s “Learning How To Learn” mentioned, but things like concept maps, activbely created and interactively shared, provide another modality I’ve found very useful for many years. Overall I think it’s not a reasonable expectation for us to remember everything we read in our minds; we need an integrated exocortex, also in writing, that captures what we remember explicitly.

It seems inevitable that automating change verification is going to be necessary, and like other previous trends like automated testing, it’s going to take us all a few years to get our heads around it and work out how to do it well.

LOL thanks for the feedback :)

Across our ~600 engineers adoption of AI coding tools is about 85 percent, and there is an undeniable uptrend in activity as a result. Parking for a moment if there is any value to that increase in activity, the question is about what benefits are there beyond that. We have found a couple of things that have caused us to change the way we do things: - by using AI supported and facilitated team alignment practices as per some of the spec-driven development ideas but separate from the coding parts, we can get teams to traction on a new initiative in about half the time, from about 4 weeks to less than two. That might not sound impressive, but initiatives typically sit around 20 weeks long, so 2 weeks is a ten percent speed up. - by taking pieces of work that are relatively shallow but need to be done in a large number of systems, we have stopped farming those asks out to the teams and started centralising those with the highly leveraged application of AI to make the changes and raise PRs on behalf of teams that they review and merge. This applies to things like resource tagging and new language translation support for new markets. these have worked well and we'll be doing more of it. In a world where 3 to 4 PRs is the norm per developer, those doing this kind of work are doing hundreds.

I'm quite sure it won't be long until tools of this kind are a normal part of every developer's environment who is coding with AI agents, because we are all burning so many tokens re-establishing our agent's understanding of our code bases with every feature we add. My current issue is finding the time to test and evaluate all the contenders but this one is going on my list as it looks extremely thorough and immediately useful in my context where I have many code bases with which I am unfamiliar but have to get a useful understanding of how they all fit together.

Concrete and practical advice for dealing with a widely shared issue in large software environments with ISO and other certifications to cater to. Thanks for sharing an article about modern AI assisted engineering without the slightest hint of catastrophising; it’s a breath of fresh air!

This is a nice frame for thinking about the divergence in attitudes towards coding with AI assistance in different environments; thanks for sharing. As someone who’s spent decades in brown field environments, my starting assumption when starting in a new environment or project is that the code will be anything but ideal, with historical layers of sediment from different eras of trends evident. I expect it to be poorly and sparsely tested, inconsistently patterned, and unpredictable to change. Most old coding curmudgeons I speak to share these expectations. What I don’t understand is why more people don’t see AI as the greatest assistant in establishing your understanding of an unfamiliar system, and why they think the code AI is writing is going to be worse than all the code we have to deal with that was written by humans in the decades up to now. Because AI is also the best assistant I’ve ever had in both establishing understanding and making leveraged changes to such code bases. But you can’t just prompt your way there expecting code changes from every prompt; you have to pants your way to a minimum viable understanding, then plot your way to well planned changes using intermediate steps such as those encouraged by the Matt Pocock skills.

Thanks for sharing an honest and balanced take after going down a path that many people would be wondering about. In the last six months I’ve built a product from scratch that I’m in the process of commercialising and I have also found a renewed appreciation of all the things that aren’t the coding, the latest being the hoops I need to jump through to get satisfractory app signing going for the Windows version of my product as a Mac-based developer. Your issues with customers not liking what you give them after a wrong assumption gets baked in rings true to me, and I’ve gained a lot more confidence using some of the spec-driven development tools and as of this week, Matt Pocock’s new /wayfinder skill to make sure the path through the fog is clear enough to proceed before starting. highly recommended.

It has always mattered for a team to be on the same page about what words mean in your domain. In 2026 it matters even more, because your newest team members are AI coding agents — and they need that agreement in an even more accessible form than humans do.

Concepticon stores everything as plain-text propositions you can read, edit, and diff. One set of propositions; as many diagrams as you need. The propositions are dense enough to serve a human reading the visualisation, and dense enough to serve an AI agent reading the text. No cloud, no proprietary format, no lock-in. Your domain knowledge is yours.

Since the late 2000s, in collaboration with friends like Simon Harris and Sebastian von Conrad, I've built various versions of this — culminating in conceptmaps.io, which I shut down in mid-2024 after the Twitter sign-in broke and I no longer had the technical skill or time to fix it. AI coding tools have now made it possible to rebuild it better than ever in my spare time.

The free Reader is for viewing, navigating and sharing concept maps — no purchase, no account, no email. Concepticon Pro ($50, once) is for the people who make them: in-app proposition editing, AI map generation, document and URL import. This first release is macOS on Apple Silicon; Windows and Linux to follow.

This is an easy to understand and apply approach, thanks for sharing. To go deeper, I might apply the thinking processes from the Theory of Constraints. As Goldratt used to say, “there’s an assumption under every arrow” in the diagrams you generate to map your logic, and uncovering them using the defined categories of objections in that process also help with facilitation among opinionated people who don’t initially see eye to eye.