Writing is thinking

Preface

This is something I wrote about internally at Trineo. Much of the below comes from a post that made its way onto the external facing company blog, but revised after thinking about what it means in the years since then, not least after the arrival of generative AI.

Value

Writing is hard but creating a writing culture is incredibly valuable. This goes double or quadruple for remote teams but any company will run better when things are written down.

This isn’t a new idea. A quick search will turn up lots of posts but I first came across it in this post by Steven Sinofsky. I’ll be pulling some quotes from there and it’s well worth reading in full.

The appeal

Writing is Thinking was an idea that appealed to me at face value. I like writing, and do find it bringing a level of clarity to ideas you can’t develop to the same extent without it.

But while I do find it enjoyable, a kind of flow state activity like writing code, I don’t think writing well, to be consumed by others, is easy. And it can definitely feel time consuming. That’s the obvious answer to why we don’t always do it. Sinofsky doesn’t shy away from this point either.

Writing Is Hard

Writing is super hard. It takes more time to write than it does to talk. It also takes more time to write a page of text than a single slide.

This is one of those things that is ‘obvious’ but easily missed how important it is. Once you see it, you’ll realise how often the easy path of slides, or even easier path of just talking, takes place - including by yourself, not just others. I’ll pull the exact example Steven used here comparing a ‘Slide’ to a ‘Story’:

Slide

Story

It’s clear how much detail is missing from the Slide. If that’s all you share, two things happen:

1. The original details are lost, forever.
2. People will make up those missing details themselves.

Think of this as a team trying to join in a lesson. Think about trying to share this lesson multiple times. Think about a new team member who only has this slide.

Even if you don’t have a ‘death by PowerPoint’ culture, you might find you more than make up for it in rushed Slack messages, video calls, or in-person walk-ups - with similarly lossy results.

Now imagine that slide was about your product, and gets turned into some features. At the extreme, you could spend six months building something that address a tangental anecdote, while missing the real point.

So what is it about the written example that makes it worth spending time on?

Context

Taking the time to write allows you to provide details, which creates context, and it matters for everything from your strategy to an individual product feature being built somewhere.

The key to achieving any goal as a team is having a shared understanding, with context kept in tact, so that everyone can row in the same direction.

In practice, I think this requires the upfront investment in writing that shares the ‘raw materials’ that go into your work. Yes, they need refining into a clear articulation of what actually needs to happen - but it should also include a thorough exploration of the thinking that lead you there, so people can weigh up the same information and arrive at the same conclusion -or not, and push back before you start.

That’s real understanding, and it prevents going off the rails later.

Collaboration

[…] you can’t have plans, especially shared plans, without writing.

The 'why' behind building software emerges from the input of a number of people.

The act of writing forces a team of experts to share the details of goals—not just the what, but the why, what else was considered, the history, context.

Writing something together is obviously collaborative. Writing yourself feels like a very solo pursuit, but once you have used it to sort your own thoughts, but at work, it’s almost always done to eventually share - whether starting something new, providing a response or building on top of someone else's words, that’s a collaborative pursuit as well.

And it’s an asynchronous form of collaboration. That might be what makes it so useful, especially for working through complex problems.

Writing is more inclusive. It is easier to contribute, doesn’t reward bullies and bullshitters, and allows for contemplation.

I don’t think it’s worth glossing over the way asynchronous collaboration like this can allow more people to contribute more fully to discussions and decisions. As well as things like English skills, seniority and all the other hidden biases we don’t think influence us, writing (compared especially to meetings) levels the playing field a lot, especially for distributed teams.

It also strikes me that reading the well considered thoughts of someone else is a very easy environment in which to stop, think and understand.

Contemplation

Providing context and fuelling meaningful collaboration are good outcomes, but you won’t get either without first refining the information you’ve gathered and thoughts that you have into meaningful (to yourself and your audience) knowledge to be considered.

The act of writing, forces the author to think through all the details and steps required to share the lesson. It avoids what happens in business all the time which is “I just know” or “experience” and brings along the team and other job functions on thinking.

I find this to be incredibly true. How often have you had an idea rattling around in your head that seems totally plausible (e.g. every “It’s Uber ChatGPT for…” idea you’ve had this week) but when you sit down to write it out, all you have is two sentences that don’t make sense and a flashing cursor… No? Just me? Ok…

[…] the process of writing and sharing thoughts is clarifying AND collaborating itself

Writing is the forcing function to clarify your thoughts and find out if they’re valuable, to sift through them to see which contradict, which are unfinished and which are just plain wrong. It’s this hard work that turns ideas, snippets of conversation, notes from meetings and everywhere else from ‘data’ into ‘insight’ that are worth sharing.

Ok but you promised you’d talk about AI

One of the things I’d forgotten until I dusted this off to write about again was the emphasis on context.

Context Engineering is the new Prompt Engineering, and from my experimentation the last few months (which I’ll write about soon) is a very accurate name. Good results come from thinking about how to optimise the LLMs context window.

Obviously that’s not the same thing but it made me smile. Context is still king, and writing is more important than ever. I think the companies that already have deep cultures of writing and sharing context are going to be at a distinct advantage when adopting these new tools. Using them best looks like working in the ideal way you never quite achieved but now kind of falls out of the process because the value is even more extrinsic.

But do I even have to write anymore?

Well, do you have to think anymore?

Now that everyone aims to have their favourite flavour of generative AI do their writing for them, a new chasm is going to open. Those who never wanted to write won’t have to, and now the Slide Deck will be a copy pasted from ChatGPT.

But used well, the tools can enhance the process of doing really valuable, thought clarifying writing. The value is even higher because now you can turn around and give that to an LLM to act on. I still see a long term advantage in a world that will be awash in content but not context, or the kind of clarity you need to underpin sustained execution against of well thought out strategy.

Valuable but spiky

One last thing. I learned over the years is some people really don’t like this approach. I used to think this was a kind of obvious way to do things to everyone, but no, it’s a spiky point of view.

At times, I’ve wondered if pushing it is even worthwhile. It’s hard to do yourself and harder still to turn into a process that can co-ordinate a group to forward. A hasty, low-value meeting that at least makes a decision and puts it into action might be better than a thoughtful and better plan that’s written down but never actioned. But that’s not a trade off you need to make, you can blend the best off both.

There is also a lot of focus on allowing individuals to work in the ways they feel makes them most effective. I agree with that, and so I’ve also wondered if pushing too hard down this path is the antithesis of trying to level the playing field as described earlier, maybe it just tilts it in a different direction?

But, I’ve seen the two ways this can go, and how the default, because it is easy, isn’t the best. I would always build a company with this culture at the core. I’d work harder on the grey areas of how you don’t just get people to write, but ensure they read, collaborate, and help make decisions quickly. I’d also realise more fully that it isn’t for everyone, but that’s true of any culture. You’ll need to be prepared for that and work on it more than you think.

For those it resonates with though, it resonates a lot. I’ll leave the last word to a colleague responding to that internal blog post years ago:

Writing is, as far as I can tell, one of our greatest technologies. Thoughts that undergo non-fixed length curation. You get unlimited time to scrupulously interrogate your thoughts. Even writing this I had to pause to try work out what I'm thinking. When it really clicks, that writing is that powerful, I don’t know why we don’t do it more.