Don't think like a brain in a jar
Your body is the best thinking tool there is. AI lets us use it like never before.
We think with our bodies and communicate ideas with our bodies, and we barely think in words. If that makes you uncomfortable, then note where you feel the discomfort. You didn’t feel that idea as a word. You felt the discomfort in your body.
And if you’re nodding along, well, yeah… You got it on both counts.
Nobody teaches the body’s role in thought in school, but there’s no doubt it’s true. One simple proof: The richer and more powerful you are, the more you get to think with your body, and the more you can afford people like me to train you to communicate with your body.
Thinking and communicating physically has become a luxury product in the 20th and 21st century: I teach it to MBA students; business and law schools built in the Harvard model grade dynamic verbal debate in postgraduate classes as much as written work; CEOs have teams of people taking notes as they speak in meetings, on stage and on the road, because that is where the big ideas happen; high-priced legal firms will have high-priced junior employees whose jobs will be to take dynamic discussion in a room and turn it into judge-ready documents. If you have ever said “Sorry, I just want to check, is someone taking notes?” you are one of the lucky few who can think beyond the written word. And if you haven’t, now is your time.
The downside and cost of physical thinking is, until recently, the best ideas would never leave the brainstorming room unless someone wrote them down. And the best ideas could not travel fast without being readable on a letter, memo, newspaper or email. Since most of us can’t afford permanent note takers, we need to get at a different style of thinking, one built around silent notes, silent writing, and silent reading. It’s miraculous what this can do. Writing is so powerful that it can take words from Socrates’s lips to my inner ear over the span of millennia. That is why we learn it. But as as a thinking tool, writing is the poor person’s talking.
Here is where artificial intelligence really can change everything. It doesn’t need stillness or silence to record your ideas like a keyboard does; it can make dull text come alive as podcasts and, going the other way, turn videos into grammatical prose. AI can make our physical thinking as permanent and easy to share as written thinking a decade ago.
Still, it feels weird to claim that writing might not be the best path to thinking. Nearly every humanities department claims to be useful precisely because “critical thinking” and written essays seem to be one and the same thing.
So let me show why non-verbal communication – known as NVC – is one of the first things my business school teaches. Once you get how effective NVC is, and how and why it is taught, it’s easy to imagine how a similar process for creating ideas might be. In fact, this more physical style of thinking is historically the way most peoples’ ideas have been developed, and it’s an accident of the printing press and the keyboard that most of us don’t practice it. It’s also remarkable how many of our most original thinkers seemed to actively avoid the tools of silent thinking: Nobel-prize winners seem to love walks and talks. AI promises to bring this style of thinking back for everyone, if we let it.
COMMUNICATING IDEAS WITH YOUR BODY
Let’s say you really wanted to convince someone that your idea was better than someone else’s. You might spend a long time at the laptop thinking about what to say and then jot down a few initial ideas. You might turn some of them into great arguments. You might even find some good data to support it and an iron-clad case that could fit on a postcard.
You're ready to head into that meeting and win in a short sharp 2 minutes.
If I told you you were only 7% of the way done, you would surely be irritated. You would ask what more there could possibly be to do?
I would admit, of course, you have done enough to convince a computer. Computers get logic.
Sadly, the human brain is harder to convince than a computer, because it is connected to a body that every second broadcasts an overwhelming and emotional amount of information.
Most of those signals stem from the way people are acting and moving. That emotional impact of a person’s tone, posture, voice, and eye contact is easier to process than what they say. It is, in fact, the 93% gap between your idea being perfect and someone actually acting on it.
If you don’t believe me, then think about a time you saw a speaker read “I am excited to share with you today…” in a monotone from Powerpoint notes. Your heart sank, this was going to be a long one. The next speaker might make the flat statement “AI is growing fast” but got you so excited and intrigued, even though the words are as obvious as “dental hygiene is important.” A lot of those reactions came from what the speakers did with their bodies, not the words they had worked so hard to craft.
This is something we know intuitively, but don’t seem to realize until it gets practically taught to us.
NVC is 👌
If this is the first time you’ve heard about non-verbal communication (NVC) it’s because you haven’t been in a management-training program. Every middle-manager learns NVC at some point on their ascent, as if it was a secret that the lower ranks should never know, or a superpower in Scientology that you only can handle when you after putting in a decade of time and money. People who don’t know about NVC tend to wonder why less intelligent people win arguments with mediocre work. People who do know about it can’t believe the magic when they try it. A silence and a stance can be as powerful as a thousand words. Speaking quietly conveys power.
The first study to investigate this found that the word “Maybe” could be interpreted in amazingly different ways when matched to photos of different facial expressions and played back in different tones. The students-subjects were asked to guess if the person in the photo liked or hated someone based on how they said this mysterious declaration of “Maybe.” When looking at the results – this is way back in the 1960s – the scientists guessed that the word itself explained only 7% of the meaning people were drawing. In other words, if you say “Definitely” and “Maybe” in the same way, 93% of the time people won’t notice the difference in your intention. When my wife uses a certain tone to say “Maybe” to some wild idea for a weekend outing my children have learned to translate that body language and tone into the complex sentence “I know you are excited about this concept and it does sound great and so I don't want to discourage you, but this is less likely to happen than world peace.” Maybe has nothing to do with it. The literal definition of certainty counts less as a word than it does as an embodied gesture.
Of course, it’s an exaggeration that all words are only 7% of the received meaning when it comes to full sentences and whole speeches (although that’s exactly what a lot of people in the public-speaking training world do!), it shows how powerfully an audience receives ideas based on things far beyond words.
If you believe that people pay incredibly close and subconscious attention to how people say things, you should obviously pay attention to them when you speak.
It also makes sense that you do the same when you think!
THINKING OUT LOUD IS NO LONGER A LUXURY PRODUCT
Why is this such a shock, when it is a clear fact of life? For ten years, my MBA students have been creating perfect decks and memos because that is the image of success they were given at school. In the real world, this picture of success doesn'tt match reality. On learning about NVC they replay the tape of those moments when the perfect deck fell flat, and a confident speaker with a mediocre idea won the day. Logical argument, great writing, clear ideas only go so far in our messy world. Ideas are received by bodies before they're received by minds.
In an age of AI, it’s time to explore the opportunities for leveraging how ideas are conceived AWAY from computers. Ideas are made by bodies before they’re interpreted by minds. We all hate staring at a computer screen waiting for ideas to come. We know it doesn’t work, and asking AI to fill the blank will lead to blankness expressed articulately in words. AI can cut through writer’s block and that sense that you haven’t captured that idea quite yet by freeing our bodies to think.
We can call this process non-verbal thinking, since just as NVC trains us to look beyond words to influence people, non-verbal thinking would push against the preconception all our ideas are formed as words. To go one step further, physical thinking captures the essence of what AI is opening up, and AI allows us to build a program in physical thinking as structured as NVC is taught in schools.
WE DIDN’T ALWAYS THINK THIS WAY
Physical acts like doodling, working with post-it notes, whiteboarding are all accepted in certain brainstorming or project-management contexts. Still, even in creative agencies, these acts have never been seen as serious or as powerful as a written speech, an internal memo, a spreadsheet. Ultimately that’s because however great the post-it notes express an idea, to get 1000 people to get it, it has had to be written down somehow. Writing has been the main transmitter of ideas since the 15th century in the West (since even earlier in China). Martin Luther’s ideas are generally taken to be the first to spread internationally via silent reading and the printing press and, following that spectacularly successful spread, we have trained our best and brightest to write well and think silently if they want to influence people in life. Most undergrads leaving our top schools are masters of sitting in silence, writing. And in business the risk of speaking your mind can be as high as the benefits: we often don’t know an idea is dumb until we say it out loud, one reason why mentors in business and thought partners in academia and the arts seem crucial to success.
It’s important to remember that writing is not necessarily a natural form of thinking. It used to be incredibly effective to train people to discuss ideas to develop them and read books aloud to spread them (it’s best to think of books prior to the 15th century as being like sheet music). The printing press turned silent words into astonishingly long-range broadcasting tools. Noisy academies were no longer the image of knowledge: Socrates spoke and developed all his ideas out loud, but thanks to Plato people received them in silence. Libraries became the knowledge’s primary physical metaphor: silence please, and stay seated.
BEFORE YOU SIT DOWN AND WRITE, WALK AND TALK
Even in the age of silent ideas, however, the history of original thinking features as much walking, debating and talking as it does libraries.
Most thinkers that have shaken the planet seem to have spent at least 30% of their productive time on some form of physical thinking.
Depending on how much of his day was spent in meetings, Einstein’s day seems to have been evenly split between what I would consider physical thinking and our more common-sense image of writing.
By contrast, when I trace mediocre ideas from clients, marketing departments and leaders to their source, I find someone silently sitting at a keyboard with great hope, coming up blank, Googling for some ideas, and then dutifully writing them down and silently tweaking them into something that looks like a good proven idea, except that it is dull and unoriginal. The kind of thing that gets an A in an exam but which no one outside of a paid examiner would choose to read. The starting point should never have been the blinking cursor of a word processor, it should have been a walk. And the blinking cursor and the anxiety of removing the blank space is a bad starting point for AI to build from, too.
MAKING THE MOST OF THE BEST EFFORT
The psychologist Daniel Kahneman began his road to the Nobel prize on the streets of Jerusalem, talking with Amos Tversky. Michael Lewis’s book on their relationship contains pages and pages of them cracking each other up and coming up with better and better survey questions that would reveal just how illogical human minds really are. They did this by walking together constantly.
It’s not a coincidence to me that the people who shook the foundations of 20th century economics did so by spending more time outside of the library than in it. It is also predictable to me that their work before they met each other and walked together had until then progressed slowly and frustratingly as they pursued traditional academic research. Fresh air seems crucial to fresh ideas.
Of course, Einstein and Kahneman still needed the mental discipline to remember their ideas from their walks, and the luck to have a thought partner that could store their spontaneous ideas. There was also still a lot of writing hard graft they had to do to get their ideas down on paper– even if their ideas were sharper thanks to their walks. AI does this for us for free, simply by being able to transcribe speech and, even more importantly, by responding to our ideas while we’re on the move. The AI-driven writing will be easier and clearer because your AI-enabled physical thinking led to clearer concepts.
SO, WHAT SHOULD I DO?
You might be thinking that at this point I’m giving everyone a free pass to skip school and never write again. Sorry. Anyone who is telling you that AI allows this is tricking you into stale remixes of already tired ideas. Writing and disciplined work is still the only way to make your writing truly original and expressive. In fact, paying attention to word choice in an age of AI is going to be more like paying attention to body language when you speak publicly. As machines do more writing, we will have to ensure we never let our attention flag and let something inauthentic or off slip out, just as a repetitive subconscious hand gesture can wreck a lecture (I am disccussing myself, here, and my own shock when I saw a tape of me lecturing).
Even if the work is still there, the balance of work you need to put in to shape and transmit original ideas is shifting to more natural forms of thinking. AI has made the hidden parts of thinking both more actionable and more powerful, and focusing on them will allow you to think bigger, think more authentically, think more originally than you did before ChatGPT arrived.
AI has made these proven but hidden forms of thinking easier to exploit. Many of us don’t have a thinking partner, like Kahneman had Tversky. Many of us have too many things going on to remember to write down that one idea we had on the walk home, or the astonishing phrase a colleague dropped into conversation on a video call. Most of us don’t have an army of people transcribing everything we say and then processing it back to us as speeches and strategy plans (as CEOs do). Nearly all of us are not geniuses. Yet all of us have the power to be original thinkers, and I want to explore the practical ways that power is much closer to us when we use AI to think physically before we turn to silence, starting with how I wrote this (overly long) post.
—Robin
QUICK LINKS
15 times to use AI and 5 not to: One Useful Thing identifies that there is a whole category of activity where the effort is the point: “Writers rewrite the same page, academics revisit a theory many times. By shortcutting that struggle, no matter how frustrating, you may lose the ability to reach the vital aha moment.” You won’t get fit by letting a treadmill move at top speed as you watch.
Getting started with AI: Good enough prompting: another earlier One Useful Thing post has this gem: "One of the hardest things to realize about AI is that it is not going to get annoyed at you. You can keep asking for things and making changes and it will never stop responding." This means you can think out loud and ramble ad nauseum, and it'll not zone out or get overwhelmed. I'd add, you also don't need to feel embarrassed at appearing stupid if you float half-baked ideas, in the same way you might with your colleagues. Although I did manage to exhaust Claude 3.5 Sonnet the other day and it made me wait 3 hours before I could talk to it again! It even gave me a warning YOU HAVE ONE MORE QUERY which felt like something my exasperated Mum would say when I had asked “but why?” too many times 😂
LLMS need to be given a purpose: This TikTok from Gianluca Mauro shows how the best version of ChatGPT plays chess worse than people, while only one model – a model that ChatGPT barely promotes – plays chess really well. The reason turns out to be that this was the one model that was explicitly told to be good at chess. This shows why organisations need to prepare benchmarking tests that reflect the tasks THEY want done by the tool. AI can only do “well” if you define “good” very clearly. For writers and thinkers who want to use AI to be creative and original, your definition of good outcomes should be creative and original.
A good point:
—Lynette
HOW I WROTE
Rule #3: Your best ideas happen in the wild, have a net to catch them
This post had two births, and the second one came before Christmas break when Lynette looked at an early draft of some of these ideas and said, “Why don’t you just do a post on walking?” I decided that I’d then try to do just that. Any time I went for a walk I would speak for 10 minutes and then ask Claude to transcribe what i said and summarize it at the top, with a strict prompt forbidding it from adding to my ideas. I found the summaries often got at what I wanted to say but hadn’t realized it yet. The passage on silent reading was a consequence of this process. More broadly, every time I did a post on walking I would find I ended up coming up with a new example or case study that would be a rabbit hole to me if I had been on a computer. Luckily, my walks are usually a 15-minute walk down to the supermarket, library or gym and I would escape the rabbit hole by dictating my main take aways and then asking Claude to suggest productive research steps to pick up later. Rabbit hole escaped! The other prompt i began to use is to ask Claude, Gemini or ChatGPT for at least 7 examples of what I had been saying. This had two big benefits: First, it showed quickly where I had not been clear and pushed me to change my language until i got 7 examples that fit. Secondly, it pushed past the cliched examples that AI seems to focus on when it comes to business or political examples, and I’ve found it is a more effective way to get truly startling case studies. Finally, I sat down to write – this intro is a mix of paragraphs I pasted in, and the writing I did raw. Crucially, the paragraphs from my conversations became a lifeboat any time I was stuck. When I could not find a way to write something simply, I usually found I had spoken it out loud in a clear way while being late to a spinning class.
—Robin
 | A guest post by
|