Weeknotes #212

This week I did:

Affecting the most important measure

We’ve been doing some user research guided by a kind-of North star of ‘effective skill learning’ to understand how much of effective skill learning is contributed to by the contents of the course, the method of delivery, and the relationship with the instructor. My sense at this stage is that the relationship accounts for about 60% of the outcome, the course content about 30%, and the technology only about 10%. If I’m right this will help us ensure we make decisions that maximise the relationship element and reduce anything that damages that.

Retros & reviews

I had some interesting discussions about how piloting new technologies and operating procedures should be used to uncover issues and weak points before rolling out to a larger audience, and not have the expectation that pilots are going to run perfectly. Things going wrong in a pilot makes the pilot a success because then you can fix them so they don’t happen in real scenarios.

Blended learning

I’ve thinking about what we build next and how our concept of ‘blended learning’ works. I’m not sure that defining the ‘blend’ by channel. i.e. digital and physical is helpful in creating the learning experience. I prefer to think of us as providing a blend of synchronous and asynchronous delivery, so that young people can access a programme that is presented live (be that in person or via video) and they are access it at times suited to them (be that by watching video in the evening or working through tasks at the weekend). How this looks will come out of our understanding of what aspect (as above, content, delivery method or relationship) matters most in effective learning. So, if the relationship has the biggest impact on effective learning then we can prioritise the part of the blend that enable relationships, but if delivery method matters most then we’ll focus our efforts on improving that. We can then decide how the blend of synchronous and asynchronous apply on different levels, so should each step in the journey and module have a way of being taken synchronously and asynchronously, or are some parts only available as one or the other but overall the programme is synchronous and asynchronous?

The ethics of moderation

I’ve been writing a discussion paper on the ethics of decision-making technologies in charities. I hope to finish it this weekend and share it with our Safeguarding Board and other stakeholders to start the discussions about ethics next week. It’s a really hard thing to write but I feel like I have a responsibility to push for the ethical use of data, technology and products.

Work in progress

I started using Notion more and got my workspace set up. I add to my library, tidied my tasks list and roadmap (which used to be in Trello), and wrote this blog post in Notion (I used to use Google Docs). I’m hoping it will improve my workflow, help me collect ideas and references more effectively as i’ll have everything in one place.

I looked at using Airtable but it didn’t have any easy way to create records from sharing on my phone, and as I do quite a lot of writing, didn’t seem like the right solution.

I have my ideas database where I record ideas and concepts that I get interested in so that I can find the info I’ve previously researched easily. I’m still using the notes section of my website for more public sharing of things I find on the internet so I need to decide whether moving them to Notion will make it easier to access previous research or just over-complicate it. I also need to get better at noting my own thoughts and ideas.

Introduction to tech ethics

I went to an online workshop with philosopher Alice Thwaite on tech ethics. We talked about freedom and how it’s more important than freedom of speech, how technology amplifies speech, Foucault and how anonymity creates power, how design is a process of changing from the current state to a desired state, normative and descriptive statements, the UN declaration of Human Rights, deontological and consequentialist ethics for handling information and making decisions, creating an ethical framework and how ethical considerations should be criteria of success for the products we build. Quite a lot for two hours.

Ethics of AI & algorithms

And I went to another of Alice’s workshops on the ethics of artificial intelligence and algorithms and talked about how power is a better way to talk about AI rather than bias because it elevates the discussion from about the tech to its affects on society, theories of power and ways power can be held over others legitimately or not, and creating a target goal, model, training data and algorithm for an AI system.

Product Management and the public interest

Kathy Pham from Harvard Kennedy School convened 300 product mangers to meet online to listen to lots of three minute lightning talks on the topics of “how product is different in mission focused organizations and companies, and what public interest tech means at this point in time”. With people from the UK, US, Canada and Philippines working in all kinds of different public interest spaces from government to parenting and housing to advice. Three minutes isn’t very long (I guess unless you’re the one doing the talking) but it gave a really wide range of the different problems products people are tackling. It made me wonder, if I was going to do a talk, what would it be about?

Thought about this week:

Tech ethics, tech ethics, tech ethics

Most of my thinking this week has been about tech ethics. From two workshops with the philosopher Alice Thwaite, reading Future Ethics by Cennydd Bowles, listening to podcasts Kate O’Neill, and Rachel Coldicutt sharing some of her thoughts in answer to my long list of questions about tech ethics in the charity sector, I feel like I have lots of different perspectives that I need to figure out and fit together.

Tech ethics is a really interesting topic to learn about, and something I want to include my essay about AI, and use as the basis for a blog post about guidance for charities introducing decision-making technologies, and write a discussion paper for work, so I need to take the time to make sure I’m taking in all the stuff I’ve learned.

Posthumanism & Actor-Network Theory

Posthumanism offers an idea to redefine humans beyond the humanist ideas that were defined in the middle ages by white men (and probably contributed to lots of discrimination throughout history) and into the future of our species as we become more connected with technology.

Actor-network theory uses the principle of generalized symmetry to say that all of the elements of a network have equal agency, including the human and non-human actors such as the systems that form a network. To me this starts to form a different approach to ethics for the future.

These both seem to be quite future-looking theories with some focus on the interaction between humans and technology so they are also interesting to think about and frame some of the thinking I have for my essay about the effects AI will have on our society.

Building an accessibility business

Jonathan sent me some documents on his thoughts about strategy for A11y.space. I really enjoyed reading them and thinking about business strategy. It’s one of those complicated real world things where no matter what model you apply nothing ever fits. It’s been a while since I had to think that way and its so easy to get tied up in knots that lead to inaction. Anyway, the internet needs to become more accessible and I’m sure A11y.space can contribute to it.

Tweets from this week:

How did Uber grow so quickly?

Scott Gorlick tweeted about how Uber approached growing. One of the fascinating things is how offline all the methods they used are. It shows how the myth around internet businesses and being purely digital are so wrong.

Digital job competencies

Dan Barker tweeted “Here’s a list of competencies that people use in ‘digital’ jobs. Which of these would you say are most important for your job/area? What is missing from the list / what doesn’t fit?”

With things like ‘analytical thinking’, ‘influencing others’, and ‘objective analysis’, the list is interesting for not mentioning words like ‘digital’, ‘data’, ‘design’. Instead it focuses on the core competencies of modern work that enable people to solve complex problems in fast changing environments rather than the traditional factory-like concept of work where workers are expected to do what the manager told them (a simplistic contrast, I know, but there does seem to be qualitative differences between the modern workplace and management approach and its older version).

Let the people learn

Hermanni Hyytiälä tweeted “If organisations want to get better at what they do, then their people have to be able to learn. Working within a rigid operating model that is designed on outdated management assumptions and related structures makes it almost impossible for employees to reflect and learn.”

Like I’ve said before, you have two jobs: learn and integrate that learning into the organisation. Employee knowledge is an intellectual asset that organisations should utilise as a competitive advantage.

Limiting work in progress

Limiting the work in progress for autonomous teams is important for organisational effectiveness

When an individual working alone has too many things to do, doing everything means that everything goes slowly. They can prioritise certain things, spend more time on them at the expense of doing other things, but ultimately everything still moves slowly.

When an individual with too many things to do works with another individuals with too many things to do, the problem is compounded because each has to wait for the other to complete pieces of work before they can work on their things. Two people can talk to each other and coordinate the work and make some efficiencies but it’s still easy to see how having too many things to do impacts these individuals.

When a team has too many things to do it becomes too complex to coordinate even just two teams with simple communication and aligned agreement. People being people tend to drift out of alignment and do their own thing. Even if the teams have someone with the role of specifically coordinating the work of the teams it’s still impossible to know how long each person will take to finish their work and so people get blocked waiting for others.

When all the individuals and teams across an entire organisation have too much to do, the compounded blocking is multipled and it becomes impossible to even figure out the current state of which work is dependent on which other work and which is blocked waiting for which piece of work to be finished. Trying to plot this into the future is a task of so many unknowns and such complexity that it is beyond human comprehension.

This tells us that the problem isn’t actually to do with the work, what is being done or how long it take, the problem is that when so many variables interact in a complex system they have unpredictable effects on each other. But one effect we can all observe is that work takes longer to complete because people get blocked.

How do we make individuals, teams and organisations more effective? How do we reduce the complexity to be more effective? Reduce the work? Reduce the interactions?

Perhaps the answer really is autonomous teams with strict limits on work in progress, but then I think that organisations look at the model and think it would be more cost-effective to slice the teams the other way.