Tag: AI

Posted on

Lessons learning – field notes from the MPA conference.

Today I spent the afternoon with around a hundred members of the Major Projects Association at the MPA’s annual conference, exploring how we create a culture of learning in major projects. 

I was there to represent the Get It Right Initiative as one of their lead trainers, using one of our Lego exercises to show how errors arise, and used this to pivot into a conversation about culture.

Some things I learnt today and observed today.

The power of switching from nouns to verbs

This from an excellent talk by Dr Simon Addyman, at the Bartlett School of Sustainable Construction.

  • Organogram to organising
  • Deliverable to delivering
  • Project to projecting.

It brings projects alive and allows them to be seen as a journey of discovery and encounter. 

In my leadership workshops I often talk about how we treat complex problems as if they are complicated. This verb-based language can perhaps unlock how we make this shift to embracing complexity.

Projects as a confluence of paths

And this also from Simon Addyman. It is traditional to think of projects as having a start and an end. But if we think about it, it’s hard to pinpoint exactly when a project really started and when it really ends. Instead we can think about it as a convergence of paths: people, individuals, organisations, circumstances converging to make something happen. 

This approach is much more in line with the concept of continuous place-based design which we teach at Constructivist. 

Lessons uploaded

We were shown the tremendous power of AI for learning lessons from project data. If the knowledge is given the right architecture then the depths of projects can be plumbed for unseen lessons.

And there is double looping here. Using AI agents to sift and quality assure the process. 

And yet…

Why do we assume complexity in inevitable

There seems to be an inevitability about complexity. Our projects will inevitably become more complex. So we bring in complex AI models that are hard to understand and make things even more complex.

I can’t help thinking that all of this complexity is a substitute for elegant thinking. Sure: the living world is complex, but it is kept in balance through the tiniest feedback loops running on the minimum of energy. 

It is elegant complexity. Instead of the kind of complexity that we seem to create which relies on increasing amounts of resources – energy, materials, attention – to work with.

Which makes me worry we are heading down the wrong path. Instead, why not:

Simplify first:

  • Clarify outcomes.
  • Gain consensus.
  • Don’t start too soon.
  • Simplify the supply chain.

Do a small bit and see what happens, rather than create grand edifices the depend on the whole thing being complete to work.

And then keep it lean:

Look at what is needed to keep each part of the process flowing.

Use all our senses to see what’s going on.

Change course based on local feedback.

Get all of that working first.

Posted on

Go (notes on complexity)

My favourite board game is Go. A 19 by 19 board. White stones versus black. You win by surrounding your opponent’s stones before they surround yours. The game has just three rules, but from this simple concept a game of incredible complexity emerges. 

My early years of playing Go were frustrating: it didn’t matter what I did, I couldn’t find a way to win. And now that I am more experienced, I find it hard to teach others. I take solace therefore that while the first computer to beat a reigning chess world champion (Deep Blue versus Gary Kasparov) did it in 1997 it took another 20 years for a computer, Deep Mind to beat reining world Go champion Ke Jie. 

The reason Go is so much harder for a computer to play than Chess is the number of branching possibilities that emerge from each move. It is just not possible to play solely on the basis of the player assessing the opposite player’s best move. And therefore a much more complex dynamic emerges in the game that involves the players ability to spot patterns as much as the patterns themselves. 

I find this fascinating. In this complex situation, the players are part of the solution. Or put it another way, the solution is function of both the physical reality (the stones on the board), the players’ perception of the stones, and the players’ perception of each other’s perception of the stones. In maths terms, the solution y = f(physical world, internal world).

It highlights for me that with complex situations in which engineers (and other humans) are agents, how we show up and how everyone else is showing up has a big impact on the outcome. We are a long way from optimum answers that can be deduced from calculation.

This post originally appeared on eiffelover.com in September 2024

Posted on

Machine work

Inputs

Outputs

KPIs

Tools

Models

Performance

Quantitative analysis

Scaling up

Accelerator

Dashboard

Timesheet

Human resources 

Bottom line 

When we think of our work as the work of a machine, then is it any surprise that the incredible machines that we have built will one day starting doing it for us.

But we do ourselves a disservice if we only think of ourselves in machine terms. If we leave out empathy, care, collective knowledge, grounded understanding of place, knowing that is not describable in words, trust, passion, play… then we are not bringing our whole selves to the work we need to do. 

There are so many more ways of knowing than the knowledge we can enter into a computer. Let the computers do the computational part – they will be very good at it – and let us step into our wider intelligence as engineers (and other humans).

This blog post was inspired by Reinventing Organizations, by Frederic Laloux. 

This post originally appeared on eiffelover.com in September 2024.

Posted on

Ultra-processed information

It’s super quick to absorb. 

Cheaply available. 

It bares little resemblance to its source. 

Its ingredients can come from anywhere. 

The growers are anonymous. 

Put together using processes you don’t understand.

It is optimised for what you crave rather than what you need.

And like other ultra-processed things:

It doesn’t quench your hunger.

It’s addictive. 

Easy to binge on.

But can be strangely unsatisfying. 

But we don’t just think with our heads — we think with our whole bodies. 

We process information by moving through the world, interacting with the environment, relating to other people, remembering through different neural centres in the body. Thinking has physical and emotional dimensions alongside the cognitive that are part of how we have evolved to make sense of the world.

When we are more active seekers of information rather than passive consumers:

  • We have to seek out what we need, creating relationships with sources, with people, with places. 
  • The process takes time, which gives us time to think.
  • We give the opportunity for our full range of bodily thinking circuits to participate. 
  • The inputs require chewing on, and this gives us time to discern what need.

The process is slower but the outcome is more nourishing.

Posted on

Field notes: operating the Decision Engine

I’ve written lots of posts this week on decision-making, and that’s because I have run three rounds of The Decision Engine workshop — part three in our Critical Thinking programme. 

The Decision Engine imagines decision-making as a production line that we build and operate. A decision travels through this system — starting with how the question is framed, moving through decision criteria, weighing subjective and objective factors, and arriving (eventually) at a decision.

It’s a model I first helped develop at Think Up during our 2015 collaboration with Arup on the Conceptual Design Mastery programme. Since then, I’ve developed it to account for everything from emotional data and gut feel to AI and emergent behaviour.

But the point is not to turn decision-making into a laborious stepwise process, but rather to build critical insight into our personal and group decision-making. 

Interesting questions that have fallen out of this week’s workshops include:

Should you start with developing ideas or agreeing your decision-making criteria?

Are we deciding — or are we building the mechanism by which other people decide?

What’s the role of subjectivity, and how do we get better at working with it?

When is a good time to decide?

And how do we continuously learn from our decisions.

Plenty to chew on, including whether we could run a day-long, stand-alone course on decision-making in future. Watch this space.