My reflections on reading “Big Mind”
If you have a managerial role in an organisation or significant investment stakes in other businesses, I am going to unreservedly suggest: this is a must-read. The principles discussed in this book concern the sustainability of competitive edges of any business.
For others, this is still a highly-recommended book. It is important and influential.
This is not the easiest book or the most entertaining book to read. But that is not a comment on Mulgan’s ability to write eloquently. Unlike many other books, which often exploit the power of storytelling to make a small number of, sometimes just one, ideas, this book is like a machine gun. Rounds of ideas fire again and again in every chapter, page and paragraph. While reading the book, I often found myself pausing to digest and reflect.
For that reason, this review gives only a snapshot, highlighting a handful of the great ideas that I think are most relevant to investment professionals. It is not a head to toe summary of the entire book. I will leave that task to Mulgan himself.
Cognitive system and collective intelligence
A cognitive system performs the work of knowing, understanding, planning, problem solving and decision making. At the micro level, this can be one human brain with its system of 100 billion neurons. It can be a group of people collaborating, without or with access to machine intelligence. At its grandest level, it encompasses the whole of human civilisation and culture. In this sense, an investment organisation is a cognitive system that operates within the larger ecosystem that defines the investment industry.
Similar to the concept of general intelligence for an individual, collective intelligence defines the capacity of groups to make good decisions through a combination of human and machine capabilities.
The concept is simple. The challenge is big.
Because to develop genuine collective intelligence, it is more about integration than aggregation.
With sufficient capital, one can amass the cleverest people and smartest computers to build a business but it can still fail. Many organisations can be very smart within narrow parameters or shorter-term framing, but far less so when dealing with the bigger picture or long-term vision. What is the secret to organising and integrating the individual components of intelligence, humans and machines, so that an organisation can be smart in solving problems?
That is the key contribution of “Big Mind”.
Key elements of collective intelligence
Let’s start with some building blocks. “Big Mind” describes ten main elements of collective intelligence, which together allow for thoughts and actions to happen on a large scale:
- A model of the world – how it works, how things cause other things to happen
- Observation – noting that what we know influences what we see as well and vice versa
- Attention and focus – from an organisational perspective, this can be shown as the ability to kill off the projects that are interesting yet ultimately a distraction
- Analysis and reasoning – a key area for machine intelligence
- Creativity – the ability to imagine and design new things
- Motor coordination – the ability to act in the physical world
- Memory – both short- and long-term memory and the key challenge being to access the right memory at the right time
- Empathy – the ability to understand the world from another’s perspective
- Judgement – the ability to make decisions, both rationally and emotionally. The book emphases the role of emotions to guide us in light of scarce and contradictory information (big minds need to be matched with big hearts)
- Wisdom – the ultimate kind of judgement; more contextual than just reasoning; integrating ethics and attending to appropriateness.
What is important to recognise is that building capability in each of these elements takes energy and time. That means there are trade-offs between them. The art of orchestrating various elements is therefore about finding the sweet spot.
Too much memory leads to being trapped in the past. Too much reasoning leads to being blind to intuition and emotion. Too much creativity leads to reduced ability to act or learn. Too much focus leads to ignoring the bigger and longer-term picture and failure to spot the novel pattern.
Introducing machine intelligence to the mix helps strengthen some of the elements (e.g. memory and analysis), but it can potentially create an imbalance. It is hard to envisage machines developing empathy. They are, at least for now, weak in creativity. And one can argue that wisdom is ultimately a human trait.
Assessing the investment industry against these elements, the picture is not particularly cheerful. There are important gaps that require a collective effort to fill.
For a start, the theoretical foundation that guides the industry’s thinking and actions is weak. As a result investment decision making relies heavily on accepted and established practice (“folklore” really) which is essentially backward looking. We in the Institute have long argued that a better model of the world is one of investment as a complex ecosystem.
Regarding observation, there is a bias towards what is easy to observe – e.g. short-term investment performance – and little effort spent on observing what is critical in driving long-term outcomes. The examples of the latter include the deterioration or improvement regarding the competitive edge of the investee companies or the strength of investment governance or culture.
Similar weaknesses are evident in the rest of Mulgan’s list: too much attention is given to managing the short term even though long-term success requires a different mindset; innovation and creativity is almost exclusively applied to product proliferation instead of solutions that align with end savers’ best interests.
Another key concept raised by Mulgan is that of infrastructure. Clever people and smart machines depend on infrastructures, both physical and virtual, to coordinate and collaborate on a large scale.
One vital infrastructure is a set of agreed rules of standards – the common language. For example, the CFA Institute has come to play an important role in setting and refining common standards for the investment industry.
However, some elements of this common language lead to distorted use of our collective intelligence. Benchmarks can produce an obsession with relative returns and a short-term focus. Value at risk (VaR) reduces a multi-faceted concept to a single dimension. The taxonomy of asset class masks the true drivers of return and sources of risk. And the concept of alpha is probably our biggest enemy: it diverts intellectual resources to competitive fields rather than cooperative ones.
Given the elements and supporting infrastructure, how does an organisation cultivate the development of collective intelligence?
We can start with autonomy. This is about how much the elements of intelligence are allowed to develop freely so they are not subordinated to ego, hierarchy, assumption or ownership. It is about allowing arguments to grow and become more refined. It is about seeking out alternative views / assumptions / models and counterfactuals as a way to sharpen understanding.
There also needs to be a balanced use of all elements of collective intelligence which I have already alluded to earlier.
Third, the organisation needs to master reflexive learning. “Big Mind” talks about three loops of learning:
- First loop: begin with models of how the world works; observe what the world does; adjust our actions and the details of our models in response to the data, within an existing framework. This loop of learning is largely reactive
- Second loop: there are too many surprises; our current models no longer work; now we need new categories and models to think with; this loop also involves reflecting on goals and purposes and is proactive
- Third loop: systematise new way of thinking; at its grandest it may involve the creation of a new field of science.
Circulating back to the investment industry, most of our learning focuses on the first loop. It is my view that we seem to spend a lot of efforts in refining models that are fundamentally broken while not working hard enough on learning beyond this loop.
This would be less of a problem if the environment were stable. In a stable environment, it is all about exploitation – making best use of the winning formula. But in an evolving world, exploration (second and third loop) is needed to survive. We sometimes have to take risks, and accept failures even when our current models appear to be working. Building redundancy into the system instead of trying to achieve optimal utilisation of existing resources is the key to adaptability.
Last but not the least, collective intelligence is about turning decisions into actions. As Mulgan puts it, it is not enough to think great thoughts and host glorious arguments. Life depends on action.
Looking beyond “Big Mind”, I will close this extended review with a few principles borrowed from another brilliant book – “Superforecasting” by Philip Tetlock and Dan Gardner.
Master Bayesian belief updating. Skilful belief updating requires extracting subtle signals from noisy information flows by incrementally adjusting probability e.g. moving from probabilities of say 40% to 45%.
Study past errors / successes. Conduct post-mortems to understand what exactly went wrong. Strike the balance between learning too little from failure (e.g. overlooking flaws in basic assumptions) and learning too much (sometimes bad outcomes really are just bad luck). And conduct project review on successes too as a good outcome does not necessarily mean there are no lessons to learn.
Bring out the best in others. This relates to mastering the art of team management:
- Perspective taking: understanding the arguments of others so well that you can reproduced them
- Precision questioning: helping others to clarify their arguments so they are not misunderstood
- Constructive confrontation: learning to disagree without being disagreeable
- Social perceptiveness: reading between the lines.
This review by no means does full justice to all the great ideas from “Big Mind”. There is, for example, a very interesting chapter on meetings in which Mulgan makes a bold prediction: soon we will use computer facilitators to regulate time, ensure everyone has a chance to speak and even suggest strategies to overcome impasses and monitor emotions!
I encourage you to read the book.
If there is one key takeaway, it is that collective intelligence can be improved, even though it’s more of an art than a science. The best approach is discovered through trial-and-error, and constantly evolves with the environment. Mulgan argues that the finance industry (among others) has failed in this iterative shuffling process and therefore become locked into configurations that keep it less effective than it should be.
We need to change this. And we can, especially if the industry takes to heart the many insights scattered throughout this important book.
 This is by no means a criticism on this style. As long as the ideas are great, it can be very effective in influencing people’s thinking.
 Here is the link to the video stream of an evening reception at Nesta, where Mulgan spoke about the book’s key themes.
 “What is a cognitive system?”, Gavan Lintern, 2007
 “Folklore of Finance”, State Street, 2014
 “Stronger investment theory”, Thinking Ahead Institute, 2016