Category: Mental Models

Saturday, 4.43pm

Sheffield, U.K.

In my Ph.D. thesis, written in 1989, I discussed the fact that when a civilization develops the technology to prevent catastrophic asteroid impacts, it marks a significant moment in the evolution of the planet. – David Grinspoon

August 2022.

That is the only month since June 2017 when I didn’t post a single article on this blog in a month.

It’s hard to remember what I’ve written in that time.

Reading old posts can be surprising, it’s like coming across a new writer for the first time.

And then you realise it’s you.

I’m feeling a bit like that as I tackle my PhD thesis again.

I had this idea that writing a thesis is a bit like plaiting fog, trying to get these ideas into some kind of order and make sense.

But it’s not really like that.

I don’t think there’s a right way to do this work.

Some fields have conventions they expect you to follow.

In the hard sciences you have a hypothesis and you test it and if your results support your hypothesis, when you’ve got something to talk about. If what you’ve discovered is sufficiently novel, then you’re good to go.

But outside that little bubble of “proper” science it all gets a little messy.

Actually, that’s not true either.

Whenever you write something you go through three phases.

First, you write a draft to discover what you think.

This is usually not very good.

Second, you rewrite your draft to get others to understand what you think.

This is much better.

Third, you rewrite your draft so it’s acceptable to the people in power.

This one is usually shit.

I mean, you have to play the game but this is why a good manager will hire a good sales writer and then use what they write without trying to make changes.

The writer is writing to sell off the page.

The manager is editing so that their manager will not be unhappy.

So they add extra words and clarifications and technical terms and generally ruin everything.

I think this happens with academic writing as well.

You write something.

Then you rewrite it to make the reviewers happy.

This is not a bad thing. The point of peer review is to make your paper stronger, and in many cases the feedback you get does make it stronger.

But, the nature of the academic publishing system means that you are probably going to have to pander to power to get your work published.

Everyone who publishes in academia is pretty open about that.

So, here’s my recipe for writing your next thesis.

First, write to understand.

Second, write to be understood.

Third, try not to screw it up.

Cheers,

Karthik Suresh

Saturday, 8.24pm

Sheffield, U.K.

Supposedly, some writers work in rowdy coffee shops or compose whole novels to Megadeth, but when I write, I wear a pair of chainsaw operator’s earmuffs. – Anthony Doerr

I don’t know what you think about LLMs and this new breed of AI tools.

Some people that I follow on social media hate it. Viscerally. They see it as a large scale theft of intellectual property made in broad daylight. Piracy. Simple as that.

There is a specialised world where that is perhaps not quite as straightforward.

The world of Free/Libre/Open Source Software (FLOSS) has put code into the world for years, because of the stubborn belief of a few people that code was too important, something that mattered too much, to be locked away.

It deserved, it needed to be shared.

If you take this massive dataset of open, working code, and then develop the tools to generate new code is that ok?

Especially when many of those tools themselves are open source?

So, we’re not losing control, no one is taking over, we still have the ability to benefit from these models while having control because of the freedoms that come with the code.

Now, I’m not really arguing the legalities or opinions on the matter – I’m not really that interested in those.

I’m interested in the practical matter of what this means for you and me.

Many years ago I had a car, a Citroen ZX. I was doing a mechanics course at college as a evening thing – I wanted to learn how to fix cars. And I learned something about the brakes.

This was a long time ago, so I don’t remember much, but the essence of it was that you needed a seal to keep the brakes working properly.

I went into a garage and asked if they stocked a brake repair kit – the seals and things that were required.

The owner looked me up and down and told me that no one repaired brakes any more. It took too long and wasn’t worth the money. You just swapped them out and put a new unit in.

He said if I was ever looking for a job he’d be happy to take me on.

I guess being interested in the details of something like that was indicative of a worker that might be useful.

Instead, I spend a lot of time with computers getting them to do what I want them to do.

You can look at code in the same way as I looked at that brake repair job – we want to do everything from scratch and do the fix properly.

Coding was, until last year, one of the few hand crafts – something where every line was tapped out by a human.

Now, that isn’t the case anymore.

Last year I wrote something. I made a thing. And, I kept a little journal of how I got on.

It took me from the 23rd of November to the 13th of December to work out the kinks and get something working.

What’s that, 20 days?

Recently, I did this job again. Now, I knew what I wanted to do, but I was coding everything from scratch.

Or rather, I was getting an LLM to code what I wanted.

I was thinking about the structure of the code and then filling in the bits I needed using ChatGPT.

And I got this version working in 2 days.

And over the last 12 months I’ve used ChatGPT to help me create analysis work in R that would have probably taken me months to do if I was starting from scratch.

So, I’m starting to think of this as the shift from one level of coding, where it’s handcraft, to the next level, where it’s working at the level of blocks or units, and it’s not important to create every bit yourself.

You still need to know how to write it if you needed to, and understand what’s going on so you can debug it and all that.

But… you just don’t need to write every last line of code.

It’s like going from using a saw to a chainsaw.

A chainsaw for your mind.

I’m the kind of person that likes the idea of old tools, I’ve written about using dip pens, for heaven’s sake.

But an 18 day saving in time is something else too.

Now, where will this take us?

One thing is that we will not need as many people doing the jobs that are there now – you’ll need fewer coding jobs in these big firms because the coders will jobs will be more productive.

Of course, the world runs on code, so you probably need more people who can do that, so you might end up with many, many more coders and a huge increase in productivity, and the ability to use all this capability to transform our economies into lean, green ones.

That could happen too.

And some people will suffer.

If you know anything about India, you’ll have heard about Gandhi and his spinning wheel.

His idea was one of self-reliance, of making your own cloth rather than buying expensive, imported cloth.

If you use FLOSS software you’re using a spinning wheel, not going to a mall.

FLOSS software makes you self reliant, but only because so many others out there have put in the time and effort to make that possible.

So, if you have a spinning wheel, what can you do with it?

I think that if you’re a person that works in an industry affected by these models – coding for starter, but perhaps writing, or music, or design – if might be wise to consider what you’re competing against now.

I’ve heard sayings on the line of countries don’t fight, economies do. Or, companies don’t compete, supply chains do.

I think in our world it’s going to be people don’t create value, workflows do.

Something on those lines.

It’s less about you and a job and more about the workflow you manage and what the results are.

If you can get a set of tools together that creates an outcome that is valuable, quicker, cheaper and better than the competition, then you’ve got something to sell.

That’s not going to change.

These tools will change jobs, but they won’t change humanity, and the way we interact.

We just need to figure out how to go to work now.

Cheers,

Karthik Suresh

Friday, 6.22am

Toronto, Canada

Let’s say intelligence is your ability to compose poetry, symphonies, do art, math and science. Chimps can’t do any of that, yet we share 99 percent DNA. Everything that we are, that distinguishes us from chimps, emerges from that one-percent difference. – Neil deGrasse Tyson

A 1999 paper by Steven Spear and H. Kent Bowen set out to decode the DNA of the Toyota production system – and seek out what made it work so well and at the same time so hard to imitate.

They argued that there were four principles, four rules, that grew over time that even the workers themselves could not articulate – they were just built into their way of working.

Let’s have a look at what they are and how we might use them as we continue to consider Lean Service design.

1. Specify work

Let’s start with what work and isn’t.

Certain things are not work.

I have a post about this somewhere and I can’t remember this exactly but I think the essence of it was that work is something you wouldn’t do if you weren’t compensated.

If you’d do it anyway, it’s not work.

Why would this definition be helpful?

Well, let’s limit it to the context of you’re a leader in a firm, someone who has a wide, creative role, who works with customers and develops services and enjoys the intellectual challenge of bringing something into the world.

You may be very well compensated for this.

But there is something that you don’t want to do. Something you need someone else to do. Something you are willing to pay someone else to do.

That’s work.

As you get closer and closer to the front line of where value meets the customer, figuring out what work is becomes easier and easier.

And this is when you must fully specify what the work is that you want someone to do.

What this means is systematically setting out, step by step. what needs to be done.

There are a number of approaches to do this, from Toyota’s own approaches to Crawford Slips, but they key is that there is a written specification of the work.

Now, when you ask for more resources and someone asks “for what work?”, you can provide your specification and say this is the work they need to do.

Most people will find this off putting, perhaps overwhelming, and file it in the too hard category.

2. Ask and you will get

The second principle is that every interaction is between a customer and supplier.

This is about a role, not about actual customers and suppliers.

In one case, your boss may be the customer and you the supplier.

In another instance, you may be the customer and the boss the supplier.

The point is that in each case, one person asks for something clear and unambiguous and the other provides something that is correct and complete.

This is about working on communication.

3. Follow the yellow brick road

The third principle is that of the path, each product or service should move along a simple or clear path.

First one thing happens, then the next and then the next.

Each action is a single piece of work that either does the work or puts it clean into flow so the next job can be done.

This is the service design work, working on getting each step working as well as possible.

4. Improvements done by the people doing the work

Train your teams and support them in making what they do better as they aim for continuous improvement.

All too often “improvements” are done elsewhere, in silos, and then imposed on people doing the work who then ignore it or find that it doesn’t do what they need it to do.

I have made this mistake and I need to learn when to recognize it – it’s hard to build something to make something easier for someone else.

You have to learn to see the world from their point of view and then start to work.

But if you’re the person doing the work you have an incentive to make it work better, more reliably, create less stress for you.

But you may not know how, so you need to pull support and guidance to get this done.

The takeaway

If you can’t do all this don’t be surprised.

The paper starts with noting that the approach is very hard to imitate.

It takes work to make the system work to do better work.

And Totyota has been working on this for decades.

One of the things you learn when you are exposed to critical thinking is that there are multiple ideas out there rather than one true idea.

Your job is to consider these ideas and weave together the ones that make sense to you.

Then you can apply what you’ve learned to your particular situation which is unique and different and requires careful thinking.

Or you can decide it’s all too hard and use some other form of management like command and control.

This approach was born in a factory and addresses products and services in that context of factory work.

There are other contexts like data work, or research work, or construction work, where you will need to develop approaches that are based on the principles and see how you do.

Now, I’m going to go and think about how to apply these principles in the work that I need to get done.

Cheers,

Karthik Suresh

Wednesday, 4.13am

Toronto, Canada

Peace is not the absence of conflict but the presence of creative alternatives for responding to conflict – alternatives to passive or aggressive responses, alternatives to violence. – Dorothy Thompson

As I consider the ideas associated with Lean Thinking I am struck by the bias towards a logical and machine-like approach which addresses but doesn’t embrace the complexities of real-world conditions.

I read somewhere that from the eighteenth century onwards the tech bro equivalents were lawyers – human society was becoming more complicated and brute power and hierarchy weren’t enough to manage evolving social dynamics.

The answer was contracting – agreeing between people what they would do – and putting a system in place that would enforce such agreements. Of course, that had been in place earlier. As Terry Pratchett observed, rich people found it was much easier to accumulate wealth and protect their hoard with the point of a pen rather than the point of a sword.

Most situations that have even a whiff of difficulty have social contracts at the heart of them, written or unwritten. The breakdown of an employee-employer relationship, for example, has as much to do with a breach of the implicit contract they have as any overtly agreed contract, which will always be subject to power imbalances.

Conflict is everywhere these days. Many countries are going through instability and pain, and responding with tribal reactions of fear and anger. Human beings cannot remove these instincts from themselves, and the inevitable result is conflict.

Unless we have tools to avoid a situation escalating to that point.

Coincidentally, there is something big taking place in a rather important country and CNN has some advice on conflict resolution, that we can apply to a range of other situations as well.

There are six points to consider.

1. Compromise or Victory?

Colin Eden’s SODA method is based on a two poles theory. Each thing you think about can be expressed in terms of two opposites. Articulating them this way helps you to see the range of what’s possible and try and see where you are on that line.

So the first thing you have to consider is whether you want victory rather than compromise. An all out push for victory is going to lead in one direction. All out compromise will lead to another. And there are plenty of examples of both leading to hubris.

The parties involved have to feel out where the situation is between these two extremes.

2. Trust or logic?

The potential for compromise can only exist if there is a chance of building trust. Logic does not work, few people are persuaded by logic. For example, you may subscribe to a philosophy that asserts that it is perfectly acceptable to act as if you are acting in a trustworthy way because the outcome you want is victory, and therefore lying is a logical approach to take if you have to in that situation. Get the other person to trust you and let down their guard and then go ahead and …

In turn, I logically know that regardless of how you act I cannot trust you.

We build trust over time, through conversation and trust-building actions. Again, it’s complicated. Sometimes you can see it in action, as when someone invades someone else’s territory. Sometimes it’s harder to spot, as when someone intrudes on someone else’s turf accidentally on purpose.

3. Gray or black and white?

Nothing is ever simple and clear and black and white.

I don’t need to go on about this, you know that when you look closely at anything it becomes more complicated and confusing.

There’s always something that led to something that led to where you are now.

And tracing it back can unearth generational trauma, passed on time after time.

Either you’re prepared to dig in and understand the layers or you’re not and want to just look at the surface level – and that decision will have an impact on how you approach the situation of conflict.

4. First move or big move?

Few things are better done fast, especially when it comes to important things. Everyone wants to move ahead quickly, and make a big change. But real change grinds slowly and you have to decide what you’re willing to do in the situation.

In the UK, for example, the Liberal Democrats, excited by their status as a junior partner in a Coalition government, naively reneged on their campaign promises and put up tuition fees for students.

A big move.

And they were nearly obliterated as a party when it came to the next election.

They should have stalled on that one – refused to budge without the support of their voting base.

Big moves require big agreement – and that takes time to build through a series of small moves.

The Paris Climate agreement in 2015, was the result of decades of slow, patient work building the science-based argument for action.

5. Dialogue or debate?

This one overlaps with some of the other points – dialogue is about exchanging ideas while debate is about winning, dialogue takes time while debate is about making your points and moving to a vote, dialogue is about consensus building, debate is about gaining power.

Which one do you think will make a situation better?

6. Flexible or rigid?

This one comes down to you and your personality. I’m flexible, perhaps too much so. I have friends who are rigid, perhaps too much so. And the potential for a good compromise is somewhere in the middle.

If I flex too much I will sign up to something that I will be unhappy with in the future, and I have to learn not to do that. If my friend is too rigid then no one will work with them and that’s not good either.

The takeaway

These points are thought-provoking and important and relevant whether you’re running a country or managing a home renovation, running a non-profit or a global business.

When two humans encounter each other there is the potential for conflict.

The fact that we have survived is because we have learned to manage it.

Cheers,

Karthik Suresh