The risk of course lies in the extremely high failure rate that all new businesses face; put up against the slim probability of potentially enormous rewards, as any tech billionaire can attest.

Venturing out on one’s own also comes with working hours that most people find unacceptable, almost zero job security (given the high failure rate), and the prospect of not drawing a salary for an indefinite amount of time.

In fact, most entrepreneurs pay their living expenses and fund their businesses from their savings. All of which constitute the payment that all long convexity bets require.

Of course, the degree of convexity within entrepreneurship also differs. This is often the result of scalability, where business models which are more easily scaled up come with higher degrees of convexity.

This is easily observed by comparing two two different businesses; a traditional restaurant in a city, versus a tech company selling software over the internet.

While both face steep startup costs, the software company is much more scalable than the restaurant, since they only have to produce their product once (although they do have to constantly update it).

Moreover, their sales are unconstrained by geography and time; that is they can sell their product to anyone in the world at any time.

The restaurant, on the other hand, has to create its products from scratch multiple times a day.

Furthermore, their sales are capped by the maximum number of maximum occupants that can be seated in their physical space, or in this age of deliveries, by the maximum productive capacity of its kitchen.

Its sales are also constrained by geography and time, and the restaurant is stuck selling food to customers from a nearby locale at certain hours of the day. The only way for a traditional restaurant to scale is to open more outlets, which still face the same set of constraints as the first.

As such, the restaurant business is inherently less scalable than the software one, with less upside potential, and hence convexity, for its founders.

This is in some sense balanced out by the much higher product development and marketing costs faced by the software company. Since being able to deliver their product anytime and anywhere necessarily means that they are competing globally, not locally.

Consequently, a job and/or career is more than the collection of factors that we make it out to be.

These would be the standard criteria which people normally use to assess their job/career options, for e.g, money, vacation days, bonus sizes, proximity to home, degree of personal fulfilment, etc.

All of these come with an element of risk that often goes unappreciated.

How secure is a job relative to its degree of cyclicality? Are the standard criteria enough to compensate for this cyclicality?

Are the potential rewards, tangible or otherwise, enough to compensate for the risks of your chosen job or career?

Will the job or career path equip you with skills that are relevant outside of your current role, so that you can remain employable if the unthinkable happens and even the most secure job becomes insecure?

Will it provide enough in remuneration for you to build up a large enough buffer of savings to guard against the inevitable economically devastating rare event?

All of these are questions that pertain to the effects of convexity on how we choose to make a living in this world, and need to be considered alongside everything else when we think about such matters.

Whether we like it or not, convexity is a part of our lives, and lurks unseen behind many decisions that we make in life – even something as basic as the kind of work we choose to do.

We cannot avoid convexity and its effects, but we can learn about them and make our decisions with a full understanding of the risks we are taking, and what we are accepting in exchange.