Month: December 2024

Saturday, 2.15pm

Sheffield, U.K.

Statistics are used much like a drunk uses a lamppost: for support, not illumination. – Vin Scully

It’s normal, I suppose, to come to the end of a year and look back to see what has happened and what has changed.

I’ve written more this year, around 60,000 words. Far short of the 280,000 I clocked in 2020, but a better showing than the last couple of years.

I also have this sense of being buried under material. Notebooks full of stuff, notes from everywhere, from stuff I’ve done, stuff I’ve read, journals that chronicle the mundane everydays.

How do writers make sense of all their material? How do they work through these ideas and get them into a form that says something useful?

My favourite author, Robert Pirsig, gives us a sense of this in a rare talk, as he describes writing his book Zen and the art of motorcycle maintenance. How the book was something he had to write. How he wrote a draft. Hated it. Put it away for a couple of years. Then wrote it again – and how this time, it came out exactly right.

The sequence that one goes through, the germ of an idea, the flailing around in the darkness, the collecting of ideas and thoughts, trying to piece them together, failing, waiting, then starting again and making sense – that’s something that we go through as humans.

Will these new tools we have – the AI assistants – help us do this better or will they make us less capable of putting in the time and work needed to go through this process?

After all, if I can jot down some notes, or copy what others have written into a file, and get the AI to group and summarise what’s going on, isn’t that the same thing that I’ve spent all this time doing?

Probably.

I think that we’ll increasingly hand over stuff that isn’t worth doing to these tools.

Reading and summarising a whole canon of ideas – maybe that’s something we leave to the AI.

Although, we don’t really need it – that’s what encyclopedias have always done. Or the introduction and literature review of a decent paper. That’s going to have the same kind of material.

The work we’ve got to do is the stuff that hasn’t already been done, or that can’t be done because there isn’t enough data to build a statistical model that can fit the existing data and predict what comes next.

If what you do can be reduced to statistics then the machines will do those faster and better over time.

Maybe that’s helpful.

What they won’t do is the stuff that can’t be statistically modelled.

I learned a decade ago that sustainable competitive advantage comes from rare, valuable, inimitable capabilities that you have the organisational structure to deliver.

I think we might need to add unpredictable to this list.

VRIOU.

Cheers,

Karthik Suresh

I often tell my students not to be misled by the name ‘artificial intelligence’ – there is nothing artificial about it. AI is made by humans, intended to behave by humans, and, ultimately, to impact humans’ lives and human society. – Fei-Fei Li

A remarkable number of conversations that I have are now about AI. Whether it’s on my social media feed, or in a professional capacity, everyone is talking about AI and what it means for us.

I think AI has transformed our lives in two ways, one that we don’t notice and one that we do.

We rely on AI for our day to day activities far more than you might think. Google maps uses AI. So does Spotify. As does that scheduling thing that tries to work out a time when everyone is free to talk. There are lots of tasks that are being speeded up and because we’re using the tools without thinking about them we don’t realize that AI is embedded in more and more of what we do every day.

The second way is the visible one – the models you can talk to and which help you with intellectual work – as a sort of research assistant.

That’s again been immensely useful. Want to write a python script or do an analysis in R. Here you go. It gets you going, gives you a starting point, that often works. It sometimes needs fixing and won’t always follow what you want, but it’s certainly faster than reading the docs and starting from scratch.

It’s not human. Remember that. I once asked for a script that would go through some data and pick out numbers that looked odd. I meant odd in the sense of outliers or unexpected patterns. It thought I meant numbers that weren’t even.

Going from this step to engineering an AI workflow that does something more is a little harder. I have a decent workflow for document analysis, something that comes up fairly often.

But after that we’re in a bit of a gray area – between stuff that needs thinking to stuff that needs a lot of mathematics. New models like o1 are supposed to be better at thinking type jobs but we run into the issue of validation. When your maps app gives you a route you know it’s working if you end up where you wanted to go. When your AI tool sets out a strategy you have to follow it for it to work – and we run into the human problem. If you succeed, how do you know it was because of the strategy? And if you fail, how do you know if you did the strategy right?

When you’re managing people you try and train them as best you can. Then you let them get on with the work and you try and check in, make sure they’re on the right track. Some managers micromanage, look over their worker’s shoulders and tell them what to do, but that’s like a prison warden and prisoners. Both are in prison. A good manager should be able to go and read a book knowing that the team is doing the work and it will be done right. The point of checking in is to validate what’s going on.

I think that word – validate is an important one.

We need ways to validate what AI tells us, an ability to test its outputs and treat them with some scepticism until we see outcomes that suggest we’re doing the right thing. Validation is about having a mental model that tells us what to expect from the AI.

For a long time we’ve talked about digital twins – digital models of physical processes.

I wonder if it’s time for a new approach – an analog twin.

A brain tool that can help with talking about what we need from our AI assistants and validating if they’re doing the right thing.

Something that we can understand.

Such tools exist – they are mental models such as purposeful activity models from soft systems methodology (SSM).

They’re just not very well publicised.

Maybe their day will now come.

Cheers,

Karthik Suresh

Monday, 10.49pm

Sheffield, U.K.

When in doubt, mumble; when in trouble, delegate; when in charge, ponder. – James H. Boren

I’ve been thinking about James Thurber today, the American writer and cartoonist.

He had a drawing style that was loose and fluid, and captured the essence of a scene in a few scrawled lines.

One of his books is called “The last flower: A parable in pictures” apparently his favourite.

It’s a story about war. About how it happens, how people change, how it makes things worse, how people make things better, and how war comes around again.

We seem to be living through a time with more wars, with more parts of the world affected by conflict.

Operations Research, the field I’m interested, was born out of wartime work.

Early work was about working out things like the trajectories of shells.

Our modern high tech economy is arguably the result of the military-industrial complex, and it’s support for better spears to fight with.

It’s all a little depressing.

Human beings develop new technologies to stay ahead, to be better equipped than others to survive.

It’s an evolutionary trait.

Failing to participate is preparing to go extinct.

At an individual, organisational, or national level we need to organise ourselves for survival.

First survive, then climb the pyramid – see how far you can get to being an apex predator.

That’s what superpowers aim to be.

I guess an argument could be made that what we do at work is try and survive the day.

And we try to do that by figuring out what the boss wants.

Everyone has a boss. Someone you answer to, someone that needs what you provide.

Your customer is your boss too.

You’ve got to try and keep them happy.

But the trick to doing that is to ask a next level question – what does your boss’s boss want?

That’s the thing you need to figure out and deliver if you want to get your work done and go home.

Cheers,

Karthik Suresh

Sunday, 8.59pm

Sheffield, U.K.

You don’t lead by pointing and telling people some place to go. You lead by going to that place and making a case. – Ken Kesey

I’m going to have to focus on my thesis now for a bit – I need to make some progress there.

So please bear with me while I work through some stuff that may or may not be of interest.

My area of research is called Soft Operations Research, or Soft OR.

OR is a field that uses scientific methods, tools and techniques to solve problems related to how a system is operated and find optimal solutions.

Figuring out how much resource or material needed to get a particular task done efficiently is the kind of problem where you can bring out the OR toolkit.

It works really well when you need to think about how to get things working.

It’s not as effective when you want to get people working.

That’s where the “soft” part of Soft OR comes in – a different set of tools that we can use with people to structure decisions and problems so we can do something about them.

My thesis is about a tool I’ve come up with, called Rich Notes, and I’m trying to figure out how to write about it.

One of the ways of doing this is to discuss its use in practice with case studies.

So what does a soft OR case study look like?

I picked up Peter Checkland and Jim Scholes’ “Soft Systems Methodology in Action” to find out.

SSMA has a number of case studies and is probably a good model to follow.

Chapter 6 has two studies in a product marketing function, and seemed a good starting point for me.

There are two ways you can apply soft systems methodology or SSM. One is to use SSM to do a study, and the other is to do a study that uses SSM.

There is a difference. In the first I say I’m going to use SSM and plan a study that is designed to apply it. In the second I do my work and if I come across a situation where SSM could help, I use it.

The second approach is where a lot of ad-hoc managerial applications happen, for example when I talk to colleagues or long-standing clients.

The former is when I am trying to suggest that I consult with a new client – and propose that I use SSM.

Regardless, I’ve done something. What now, how do I describe it in my thesis in a useful way?

There are six things to consider, as I’ve gleaned from Chapter 6 of SSMA.

First, describe the context – what’s the background, what’s the situation, and how did you enter it?

The thing that people are most curious about is how it all began.

It’s like asking a couple, “How did you meet?”

It’s that context that helps us situate ourselves in the situation.

Second, from an SSM perspective, it’s worth understanding what’s going to be the end result.

Sometimes it’s a report. Sometimes it’s an outcome – some kind of change for the better.

Did you know at the start what sort of end result you were aiming for, or were you making it up as you went along?

Third, how did you gain an appreciation of the situation?

This is the important bit – seeing the situation from the points of view of the people involved.

It’s not about one side of the story but getting multiple perspectives and seeing what’s going on with fresh eyes.

How did you do that?

Then we come to the last three steps, which are a bit more technical.

Fourth, what systems did you conceptualise?

A system is about parts and connections – what are the bits and pieces that make up the situation you’re studying?

Fifth, what conceptual models did you build?

A conceptual model brings the system to life.

This is probably not going to make sense unless you already know a bit about this topic, but think of it like this.

A system is like the parts of a motorbike. The ignition, the gears, the handlebars, the fuel tank, the wheels, the frame, and how they are connected.

And don’t forget, the rider is also part of the system – maybe that’s you.

The conceptual model is how you start and ride the bike.

The first is static. The second is dynamic. Together, they get you going.

Sixth, you compare your models with reality and make changes.

What does your model say should happen, what is really happening, and what needs to be done to make things better?

In your case study, what did you do?

Bring these six pieces together and that’s how to write a soft OR case study. Should be good to fill a few pages.

Now I need to go and do a few of those.

Cheers,

Karthik Suresh