Thoughts from a Thirsty Bear

As AI adoption continues apace, one good thing coming out of this second mass experiment of the decade (the first being Work From Home, caused by Covid) is  data . As more and more companies adopt AI in the software development workflow, they are generating huge amounts of information about its effectiveness. The folks at CircleCI are in a unique position to collect the data across millions (28 million to be precise) of delivery pipelines they host. They have published a summary of their findings  here . So what do 28 million data points tell us about using AI in software delivery? It's..."interesting". TL;DR - AI's effect on branches is noticeable. It increases throughput of branches for median teams significantly (15%), but this increase never makes it to production; they are never released. The production release branch throughput degrades by ~7% for the average team - AI adoption slows down teams by quite a lot. As expected, the AI-induced pressure of generating m...

 I noticed this throwaway comment recently: " Whenever I use AI to probe a topic I know something about, it seems to make a number of errors; by contrast, whenever I use it to explore topics I know little about, it knows so much more! " I must admit I have noticed something similar. When I am asking about something I know about, such as software, AI generated responses tend to be at best flawed, at worst plain wrong. Yet when I ask about something I genuinely know nothing about the results are compelling and I have to rein myself in. It turns out this is a named effect, observed by Michael Crichton (yes, that  Michael Crichton !). " Gell-Mann Amnesia ". He noticed that experts reading articles in their field of expertise will tend to find them full of errors, whereas reading other articles, even in the same publication, on topics outside their expertise  are believed to be credible. That is exactly what is happening with AI. When the results are critiqued by an exp...

What makes a high performing team tick? What is it that takes a team from being medium or low performing into elite performance as defined by DORA? On demand deployments, single figure bug counts, next to no downtime. Having worked in high performing teams, as well as having built a few myself over the years, I feel that I have a unique perspective on what it is that allows a team to work effectively together. The "Secret Sauce". So what's the recipe?  It's...." complicated ". But it looks a bit like this: At the top, unsurprisingly there is Technical Excellence . Team members need to be fully aware of and skilled in the various accepted best-of-breed tools and techniques that are out there. These currently include TDD, BDD, incremental design & architecture, refactoring, pairing/mobbing and so on. No excuses, these are all proven techniques that improve both quality and delivery, and should not be dropped unless there is a provable reason to. Even then ...

Do you know where your team's time goes? What are the competing priorities? Understanding this can be the first step to making sense of how a team is working together at a given time, and gives fairly heavy hints how to improve things for the better. A team's investment (time, money etc) can be broadly split into four 'pots': External Demand . User-facing features, new functionality. Internal Demand . Improvement. Kaizen. Failure Demand . This is the surprises. Often nasty, time consuming surprises. Business As Usual (BAU) . Keeping lights on. Those essential maintenance tasks that keep things ticking over. Any team that understands this, and knows how interdependent they are will generally thrive. Wait... interdependent?   Yes indeed. This is a zero sum game, and any team needs to make a conscious decision of how much of their finite budget (people, time, money) to invest in each at any given time. To make things even more interesting, a team only has direct control o...

 We are certainly living in interesting times right now. Specifically here I am talking about the rise of "Artificial Intelligence" (or, more correctly, 'Large Language Models", AKA 'LLMs') in software development. Whether you believe these tools are useful, or will lead to the demise of the human race¹, there is no doubt they are surrounded by the biggest hype cycle that anyone has seen before. There is an unprecedented (many are now saying unsustainable) amount of investment happening in this area. But any investment requires a return which means marketing and sales are working overtime to bring in new paying customers. Unfortunately as a result of these intensive sales campaigns there are companies - many of whom should know better - becoming ensnared in the unfulfilled promises being made by unscrupulous companies, and are " drinking the Kool Aid ". However, once the contracts are agreed by senior management, and the (not insignificant) bills be...

I often describe the way I remember things as "hook and loop", or a "Velcro mind". Over the years I have learned and unlearned so much as technology moves on, and this analogy goes some way to explaining how I am able to do this quite so effectively. To start, I am sure everyone is familiar with hook and loop temporary fastening. Also known by its "Velcro" eponymous trademark. The non-sticky sticky fastening can attach something, then be pulled away and refastened to something else. It is "hook and loop" because that is exactly how it works - on one side it has an array of tiny flexible hooks, and on the other, fluffy side the fabric forms lots of loops. Put the two together, and the hooks hook into the loops until pulled enough to unhook the fabric. So what's that got to do with minds and memory? Imagine learning a host of concepts. Background ideas, principles and models that apply to  broad range of related subjects. Sometimes people call ...