A researcher once asked me about my motivation for working in AI and whether I would have been interested in this field twenty or thirty years ago. He was prompting me to think about whether it is more worthwhile to dedicate one’s life to a single pursuit or many. Often, people aspire to build a career around their passions, only to have unforeseen realities derail their ambitions: the pram in the hallway, golden handcuffs, domesticity.

However, reality is rarely as straightforward as these phrases suggest. While a clear dream may fade, the underlying motivation finds fulfillment in unexpected ways. That curiosity takes on a new form, often without us making the connection, or catching us by surprise.

To the best of my knowledge, this behaviour explains why people become tax collectors — I’m convinced that nobody truly dreams of collecting tax as a child, but some grow up and happily turn this into their career for decades. Why? They find unexpected enjoyment in the day-to-day tasks; collecting tax lies close to a fault line they never knew they were aiming for.

In this way, the greatest career one can ask for is one where each job is ambitious and important – i.e. aligned with their common theme. By allowing ourselves to obsess over certain aspects of the job, we make few compromises, and, given enough time, self-actualise by working on deeply satisfying projects for a long time.

But this idea can also explain work that is truly revolutionary. Some academics are thrilled by the idea of divergence, generalisation, or the strength of a narrow proof. Satisfying this deeply held curiosity becomes a powerful heuristic for the best in the field to make connections that are nothing short of visionary. How else might they solve a problem at near inhuman speed, or notice some minor detail nobody else thought to? Science’s hardest problems are simple to those with a mind geared to solve them.

For instance,

• Feynman’s theme appears to be compression: exploring many things and transforming them into a form that is elegantly communicable with enthusiasm. This is a great interest for a teacher to have.
• Altman’s seems to be scaling: persistently seeking exponential returns governed only ‘by god’. This is well captured in a favourite question of his, “what’s the one thing, done today, that would help you grow fastest?”
• LeCun’s theme could be described as generalisation: through the exploration of ever more powerful learning representations. Each model he develops distinctly becomes more ambitious along this trendline.

Of course, strong performance in any field doesn’t just depend on an inate ability to navigate hard problems from a unique angle, it also requires real work. But this idea explains why moments of epiphany happen for some. A friend of mine is extremely gifted at mathematics, and recently told me that he despised maths and everything to do with it just a few years ago — but he had stumbled upon a moment of clarity, triggered by a classmate’s description of a physics problem, which compelled him to suddenly obsess.

In A Mathematician’s Apology, G. H. Hardy wrote:

I do not remember having felt, as a boy, any passion for mathematics, and such notions as I may have had of the career of a mathematician were far from noble. I thought of mathematics in terms of examinations and scholarships: I wanted to beat other boys, and this seemed to be the way in which I could do so most decisively.

But then, after reading C. Jordan’s Cours d’analyse:

I shall never forget the astonishment with which I read that remarkable work, the first inspiration for so many mathematicians of my generation, and learnt for the first time as I read it what mathematics really meant.

The final point I want to make about this idea, before I explore my own theme, is that it can help understand why some types of change are harder than others; why some changes cause grief, and others are inviting. Times are painful when we are moving away from something that attracted our deepest curiosities. This is most obvious when we feel the loss of something that bought us a sense of achievement. In contrast, time moves fast when we are engaged in activities that align with our deepest interests. Such pursuits are autotelic — intrinsically rewarding and self-fulfilling.

---

My bachelor’s degree was decided at the age of fifteen, while I was interning at the Canary Wharf Group as an aspiring architect. While there, I was invited up to Level39, where London’s fintech startups were incubated in a single shared workspace.

At 3 o’clock, a tray of cookies was brought out to “organise serendipity” between teams working on similar projects. Having carried the tray, I stuck around to listen, and was struck by the calibre of conversation.

My supervisor explained that these startups might be having discussions that pivot their entire mission in an afternoon. A tray of cookies catalysed decisions about the next features they would be launching, how they would generate revenue, or the very core of their mission. Experiencing the butterfly affect in this manner — a small nudge creating massive downstream impact — was something I remember telling everyone about that year, whether they were remotely interested or not.

A year later, I returned,and noticed that seven receptionists had become one. My supervisor had built an app in his spare time to connect founders directly to investors and one another, which had drastically increased the bandwidth between operations and had deleted a lot of the beuracracy. The receptionists had moved onto new projects where they were needed more.

A year after that, I worked at a cheese factory on the production line. We were building cardboard boxes, filling them with straw, stacking layers of cheeses and ice packs, and packaging them up with delivery stickers. Each stage was performed sequentially and in batches, yielding fifty boxes per batch, five times a day. It was fast, but felt wasteful: some of us weren’t good with the tape gun, and everyone was slow moving between fridges. One morning, I convinced my friends to switch specialise, so that each of us could practice our technique and optimise. The bandwidth was lower with only six boxes on the line at any time, but the throughput was three times higher. From then on, we were sent home at lunch time each day for the same pay.

The junctures where a small nudge can lead to strong change is something that, upon reflection, seems to be the kernel of thought that motivated my interests for the longest time. In my second year studying economics, I took part in a big hackathon in the electronics department. As anyone who has participated in a high stakes hack can attest, the first one is always sobering. I watched a team of twenty business students be demolished by a solo coder and a can of Redbull. I swore to learn his craft, which spiralled into an obsession with AI and building. My research focus began from the top-down; understanding how ML and robotics will affect the workforce, and developed into multi-agents in simulation.

Finding the one act that affects everything else is my obsession. It is also a meta-theme, hence my compulsion to share. What do your favourite ideas orbit around? When are you at your best? These questions can help you identify your common thread.

Once this idea is planted, human curiosity demands that we know our own.