Weeknotes 401

This week I did:

I’ve started so I’ll probably never finish

I started writing a few blog posts that I’ll probably never finish.

The first is about a flexible, principle-based approach to being more agile using my three word definition of agile being about ‘uncovering better ways’ – so uncovering better ways to lead, uncovering better ways to work as a team, etc.

The second is about which narrative about AI you choose and how it will change over time. I guess it’s kind of like a Lindy effect thing where the longer AI is around the harder it is to push the narrative it’s just a bubble or that it’s an existential threat to humanity. Which means that over time, the narrative will trend towards the middle of AI being new tech, just like all the old new tech. Photocopiers changed the world too, ya know.

I read/listened to:

The Double-Edged Sword Of Diversity In Teams

The Liberators Podcast is back with a really interesting study on how diversity in teams affects team productivity. Some of the insights include diversity makes teams more productive when they do work that doesn’t depend on each other but less productive if they have to work together, because having diverse people in a team leads to more conflict. And leveraging the different perspectives diverse people bring to a team doesn’t happen by accident, it require intentional effort. There’s also an interesting part about psychology safety that says, if you want diverse teams (because it’s the morally right thing to do) then you need to create the psychological safety for that team to work effectively. Psychological safety and diversity go hand-in-hand.

The coaching habit

Started reading The coaching habit to help me get better at asking questions. I heard somewhere (forget where) someone say something about only learning when they’re listening, because when they’re talking they are saying things they already know. So, as I expect the next few weeks to demand a lot of listening, I want to try to make sure I’m hearing the right things.

AI Adoption: Sparking Digital Transformation in Nonprofits and Charities

I read this article because I wondered if it might have something insightful about AI being the catalyst for digital transformation. Sadly not. It says, “AI may represent a seismic shift for nonprofits that choose to embrace it, bringing not just another new tech tool, but a radical reimagining of how they operate, engage with their communities, and fulfill their missions”. Yes, it might. But then the article goes on to say, “AI holds immense potential for nonprofits to improve organizational efficiency and effectiveness. From automating administrative tasks to personalizing donor experiences, AI can revolutionize how charities operate.” I’m not sure automating admin tasks is quite the revolution the charity sector needs.

And I thought about:

Safe Bets

The Civic AI Observatory newsletter talks about the three AI projects from the UK Cabinet Office’s new Incubator for Artificial Intelligence. As the newsletter says, the use cases for AI seem like safe bets. They are all essentially, ‘search-and-summarise’ with a ‘human in the loop’ use cases to surface information buried in documents that no one would ever have the time to organise. They are turning unstructured data into structured data (I wonder if they organise the structured data or always just go back to the unstructured data for each query).

Anyway, as much as I believe in the ‘technology push’ approach to innovation, I’d really like to see the adoption plans and metrics for the three projects. I’m particularly interested in how much instruction people need before being able to get the most out of the tools – is it entirely intuitive or do people need to know how to write good prompts? What mental models to people bring to these kinds of AI tools – do they think of it as like Google search or more like live chat with a person?

Personally, my current thinking about AI adoption is that it’ll become ubiquitous behind the scenes for things like cyber security threat detection and extremely large data set analysis, but chat won’t be the next big interface that means everyone knows they are using AI everyday. So, it’s interesting to me that the incubator team have decided to go with three user-facing products that bring adoption challenges. But I’m sure they have great product people thinking about these things and how to turn the ‘tech push’ into ‘market pull’.

You can’t service design an organisation

Saw another LinkedIn post about using Service Design to design how organisations work. I realise I’m in the minority thinking that’s a bad idea, but that’s ok, I’m used to it. If we understand organisations as complex adaptive systems, then the idea that a successful organisation can be designed up front makes no sense. Creating an effective organisation demands a very different approach, one that’s much more about setting up ecosystems and experimenting the way forward.

Canary in the coalmine

Been thinking about this analogy quite a lot, partly because it’s used in articles about autistic people at work. I’m still not convinced it’s an entirely healthy analogy when the canary refers to a person because their job is to die in order to protect others. But, I do like the idea of looking for small signals which show bigger issues. Maybe the canaries don’t have to be people. Maybe they can be things like behaviour patterns in a team that change over time or a competitor launching a new product.