Dear Reader
I’m back from the land of chocolatines (emphatically not pain au chocolats) and back to school promotions for cigarette lighters, see photo below.
Also, hiking is called Rando in France. Giving rise to magazines like:
After such simple times, we continue our explorations of complexity…
Two Numbers That Limit Our Complexity
The complexity of the current world presents two problems for all of us:
We can't handle complex problems using our brains alone.
We think we can cope.
Each of these problems is caused by two measurable limits on how we think.
First: working memory. We cannot multitask, and while doing one task, we can only keep three, maybe four, things in mind at any given moment. This means, among other things, that if the question is "What should I do next?" and there are more than four things we need to get done, we can't work out which should be first. So we use lists—flawed, weird, idiosyncratic lists of big things and little things that we need to accomplish.
Second: overconfidence + intuition. In his recent talk (see previous Antonyms) Daniel Hulme, the Chief AI Officer of WPP, introduced another limitation, echoing Daniel Kahneman’s insight in Thinking Fast & Slow: “we're confidently using our intuition every single day making the wrong decisions”.
He used the example of having a number of people and a number of jobs that needed doing. This is a challenge we call resource planning. The problem is that we can't quite understand how complex the challenge becomes very quickly:
Q: How many ways of allocating five people to five jobs?
A: 120
Q: How many ways of allocating 15 people to 15 jobs?
A: More than a trillion. (1,307,674,368,000)
Now think of your to-do list. And the number of hours in a working week. It’s a similar issue, before you even get to the point that there are too many things to do in the available time, the possible ways to prioritise and organise yourself may be overwhelming.
The answer?
The answer to complexity is technology. Which creates more complexity. And then need technology to deal with it. Right now we are victims of overwhelm caused by the last 25 years of tech innovation. The answer is… AI.
How did we get here?
Everyday life is complex, increasingly complex because of the amount of information we have or have available, the number of communications, the number of options to choose from.
The long arc of human progress for the last 10,000 years or so has been a process of co-evolution with technology. Our hardware (bones, brains, and associated sensory gear) hasn't evolved much in that time, but our technology and our societies have evolved together, spurring one another on. Both of these things have helped us rewrite our software (how we think).
Like those cautionary tales about genies and wishes, sometimes getting what you wish for can turn into a big headache. You want to be able to provide more food for your tribe, so you innovate with farming methods. No more moving around. The tribe grows—grows a lot, in fact. An unexpected consequence is that not everyone has to be farmers, and you can get specialists who create trades and technologies of their own. Hurrah!
Daniel Hulme’s company solved that mega-complexity problem of resourcing we talked about before with an algorithm for PwC, the consulting firm. It worked! It worked so well that billable hours increased massively, but also people started finding themselves working on fragments of projects and enduring more long journeys to get to jobs. Clients complained that they were seeing new people on their projects every week. Where was the continuity?.Hulme says that they’d not thought to ask: “What if it all goes right?”.
We’re mid-way through the transition from the industrial to the digital age, a process that began in the 1990s with the mass access to the Web. In the last 30 years we have created all manner of digital tools with unintended consequences without asking “what if it all goes right?”. Take some prosaic, everyday examples from most people’s working days:
Email will remove the friction of physical postal and message systems. Hurrah. Consequence: 400 emails a day.
Online diaries will mean we can all see each other’s schedules and find space for meetings that work for everyone. Consequence: So many meetings people frequently have to do long hours to get actual work done.
We can measure everything people do online to target ads more effectively. Consequence: surveillance capitalism + no one really knows what the real numbers are anyway (see below).
Breaking Things Down
We asked last week if bad delegators were also bad AI users. Actually, we're all bad delegators if it means delegating things we're good at because we have already written our own little apps in our heads to do these tasks. Little automations. Heuristics? Muscle memory?
The difficulty in explaining something you're highly skilled at stems from a phenomenon psychologists call "expert blindness" or the "curse of knowledge". Here's why it happens:
Automaticity: When you become an expert at something, many of the steps become automatic. You no longer consciously think about each individual action, making it hard to break down the process for others.
Unconscious competence: In the four stages of competence model, experts operate at the "unconscious competence" level. They perform skills without actively thinking about them, which can make it challenging to articulate the process.
Chunking: Experts tend to group multiple small steps into larger "chunks" mentally. While this is efficient for performance, it can lead to skipping over details that novices need.
Implicit knowledge: Much of an expert's knowledge is implicit - things they've internalised through experience but haven't explicitly verbalised.
Forgotten struggle: Experts often forget the difficulties they faced when learning, making it hard to empathise with beginners' struggles.
Jargon and shortcuts: Experts develop specialised vocabularies and mental shortcuts, which can be confusing to novices if not properly explained.
Intuition: Many expert decisions are based on intuition or "feel", which is notoriously difficult to explain.
Context blindness: Experts may take for granted the contextual knowledge they've accumulated, forgetting that novices lack this background.
Nobody knows what a million views looks like
I love this superb long-read on how weirdly imprecise online audience measurement is by Julia Alexander, an expert in online media analytics.
In a nutshell: Online advertising works, but no one knows what the metrics really mean.
Because of fraud and varied standards, it “is still a guessing game about whether those people are even real."
If you don’t have time for the whole thing I asked my Gen Z AI bot to summarise it for us:
So, like, 100 million views might sound super impressive, but honestly, it's just a number in a chaotic online world where nobody really knows what's hot anymore. Basically, if you're not in a niche fandom or streaming the latest TikTok dance, good luck getting any real attention—it's all just a digital popularity contest and the scores are fixed. And wrong.
Recommendations
Reading
Antarctica, by Claire Keegan
These short stories. Each a rattling ride through sucking, dark floes of others’ lives. After finishing one I rock on my feet and take a breath. Once or twice I’ve moved on to the next, but most times I've had to set the book aside for a while. Waited for my own jarred feelings to recover a kind of equilibrium before I returned to be pulled through another story.
Emotional Overdraft, by Andy Brown
If you run a small company, this is a very useful book that helps you not keep burning yourself out to keep things running. Thoughtfully written and straightforward. Just what the doctor ordered.
Transit, by Rachel Cusk
The second in Cusk’s Outline trilogy, essentially short stories connected by the constant of the author’s point of view.
Reality edition of Nautilus
We don’t usually recommend magazines in the reading section, but the current issue of Nautilus (available online) is a spectacular deep dive into ideas about reality, including:
A physicist confessing to reality-hunting
A composer exploring dark realities
A hermit's search for God
Insights into psychedelic experiences
The discovery of black holes
TV
We Hunt Together (BBC iPlayer).
A great chalk-and-cheese combination of detectives up against a capricious psychopath and her vulnerable but extremely capable lover make this a wonderfully diverting series. I’d say perfect Sunday night TV, but you’ll want to binge the whole thing. And there’s a second series waiting for you if the first hits the spot.
Succession (NowTV, but buy it).
I’ve gone again. I will keep recommending this because there are still lost people who haven’t felt its dark brilliance shining on their souls.
The Rings of Power (Prime)
I’m not sure if it’s good. But I am watching.. While I wasn’t particularly waiting for a second series, come launch day, there I am goggling at pre-Hobbity Middle Earth fireworks. The first 20 minutes of Sauron backstory was fun, and then the elves started having long conversations about “nyah-nyah-nyah” and how incredibly significant things were happening and I got very bored. Maybe it’s a vibe, or maybe it’s the spectacular CGI bonfire of the budget that brings me back. But I do enjoy it, on some levels.
That’s all for this week
Thank you for reading. If you liked it, stick a “like” on it and I’ll come up with something again next week.
Antony