How to Spend 10000 Hours of Your Life

Once you start with the premise that you live life only once, and that man’s ultimate resource of scarcity is time - what and how we spend our time working on becomes self-evidently important. The richer you get, the more that you get to spend to acquire other people’s time to do your bidding; so something I thoroughly enjoy about coding and A.I. is the fact that it allows me to leverage tireless machines to do my bidding without worrying about their feelings, how awesome is that? It’s been making me wonder about how the Stolper-Samuelson Theorem might play out for labour and capital in an age of A.I. but that’s food for thought for another time.

Having completed the Le Wagon data science bootcamp and as I work through the MITx MicroMasters programme in Data, Economics, and Development Policy, it’s clear to me that I want to spend at least 10000 hours of my life working at the intersection of economics and computation. Intellectually, they fascinate me. Practically speaking, the whole point of studying the world is to change it, hopefully through evidence, plenty of plumbing and randomised controlled trials (RCTs) rather than gut and feelings. I find that I’m less concerned about the domain of application than I am about the vertical of know-how and method because at the end of the day, the principles and applications of economics and computation work as much in art as in discrimination, immigration, trade, finance, climate change and pretty much any human endeavour of interest. The tools and technologies are already here, the future is already present, they’re just unevenly distributed.

Measuring Return on Happiness

In this sense, I am pleasantly surprised by how far the idea and practice of RCTs have come since I last read “Poor Economics” many years ago and believe that the world will be a much better place if we speak the language of data and evidence when assessing impact, rather than conduct simple pre-post evaluations, give ourselves a pat on our collective backs then call it a day. I’m curious about how the world might look like once we measure and benchmark social programmes on their Return on Happiness (RoH); if subjective well-being is already accepted as a “catch-all”, as “an aggregate way of assessing someone’s overall well-being” - why not extend this idea further?

I, A.I. 爱

The purpose of this post and subsequent explorations is to lean into economics and computation, specifically A.I., as they apply to areas of interest to me: GLAM institutions, enterprises, games and entertainment, climate and sustainability, impact, governments, you and I and our grandchildren. It is in this spirit that I gave a presentation on seeing and searching collections with machines at MuseumNext Digital Summit 2023. The sheer volume of content is staggering so what better tool than A.I. to help us to see in new ways as well as search and discover content that might be inaccessible otherwise?

ARTificially Intelligent: Seeing and Searching Collections with Machines

“My boss struggles to understand social media, how do I get them to buy into A.I.?”

There was a question raised about how do you get a boss who struggles to understand social media to buy into A.I.. It’s an interesting question to me because I myself have been wondering: how do you get adoption to move in sectors “untouched” by the competitive pressures of private industry so-to-speak? On some level, I feel like one does not even need to make an argument for A.I. in the private sector, because the pursuit of profit requires no further incentivisation for adoption. How about a sector in which cost, capability and counterfactuals could all be gapping gaps before one even lays the first brick?

There is a world in which a museum adopts A.I. and a world in which it does not. In both worlds, the default configuration is that market competition does not matter, especially if some form of state-mandated monopoly is in position. There is nothing wrong with public institutions per se, if anything, we need them precisely because there are places where the invisible hand of the market fails to reach; so my knee-jerk reaction is almost: maybe in situations like this one needs to fire your boss and go somewhere else. The crux of the matter is that oftentimes in organisations of scale, top-down leadership is absolutely essential to acquire the legitimacy and authority necessary to create change. Yet this feels like a hardly satisfying response given the importance and urgency of the issues public institutions are tackling. It feels extremely lame to rest the entirety of success of how we integrate refugees, encourage creativity and deal with misinformation, ensure that the gains to trade is evenly distributed, and so forth on one individual.

In this way, I’m intrigued by the idea of letting the competitive dynamics of measuring and calibrating the performance of non-market activities and organisations by their RoH. There is a set of treatment organisations adopting A.I. and a control group that does not. If the treatment group fares significantly better on the metrics that matter on a per dollar value basis, now we know. Now we know where our tax dollars and time should flow to.

A lot of people will balk at the idea but: why not? We constructed Gross Domestic Product (GDP) as a measure because if we didn’t even know how much output we were producing and how much it was changing, you are really shooting in the dark and making blind guesses as breadlines meander the streets of New York during the Great Depression. Maybe this is simply one possible response to the climate crisis that is worth trying. Maybe A.I. will be especially valuable in sorting through all the complexity, data, noise and sewer to give us a gauge of the data points that can predict and maximise human well-being. This is not to say anything about causality or to make a value judgment on the predictors per se. If anything many things we learn may turn out to be scarcely interesting, if at all.

So this is my first brick on a narrow set of cultural artefacts. 😎 Next up I’ll discuss intelligence as an empirical exercise, training A.I. “Stefanie Sun” to sing in Vietnamese and maybe flesh out some thoughts on the relationship between labour and capital in an age of A.I. slightly further.

Originally published on PubPub at erniesg.pubpub.org/pub/hnn0p7d9.