On proof of work and competitive industries: Ponzis, bubbles, buying high and selling low

On why many intellectual jobs are pretend jobs, why pretend work is actually a valid way of assessing real work, ponzi schemes and imposed structure.

Graduate recruiters & Team JUST

When I arrived at Cambridge in 2011 I was immediately approached by hoardes of corporates trying to sell me (a clueless teenager) a vision of what I should do after university. Law firms, investment banks and consultancies spend a large amount of money getting you (the clueless student) to go to career events. All you have to do is listen a bit to what it is they do (because you have no idea) and you’ll get some booze and food in exchange. You go there and listen to some strange corporate mumbojumbo, one of the many secular equivalents of sitting at Latin Mass. The company priest (HR person) explains the company values. You should join this company because it is different to other companies, the same way the United Methodist Church is clearly different to the Free Methodist Church. The HR person is followed up by someone that clearly has more important things to do, but is evangelising to boozed up teenagers out of a sense of duty. They are here to sell the a vision, a structure of reality.

Many years later in 2018 I sat in front of my laptop, exploring the rules of FOMO3D. FOMO3D stands for “fear of missing out in 3 dimensions”. FOMO3D is a game, but it is testament to human beings’ ingenuity: it is a self-enforcing contract on the Ethereum blockchain. In theory, nobody can stop it. With some simplification, the rules are the following: whoever buys a key such that nobody else buys a key in the 24 hours that follow, is the winner of the pot. Once you’ve bought a key, you get a paid percentage of the amount spent buying keys after yours (a dividend). You get rewards for recruiting others into the scheme. As of writing, over 40m dollars of ethereum have been paid into the contract, the pot is worth more than 10m dollars. There is no ambiguity: it isn’t creating wealth, you’re merely hoping that you won’t be the last sucker (or that you’ll be the very last sucker). It is a shameless Ponzi scheme, a pyramid game hosted on a website called exitscam.me. It was made by Team JUST (where JUST is slang for losing money). The structure of the game is clear and immutable.

Career events start much more ambiguously. Nothing much is actually said at these events, so you discuss things with fellow students, and with some helpful prodding of the HR person (priest) you come to understand what is expected from you. First of all, you need to get a short internship (I think they called it a summer insight back then) in your first year, a summer internship in your second year, and hopefully by the end of your internship you should have a job offer which is conditional on having good exam results on your last year. Its all very nice, structured and familiar. The same way you have to jump through structured hoops in school (i.e. extracurricular activities, good grades, etc. etc.) to get into a good university, you have to jump through hoops to get into a Graduate Programme at one of these companies. All your friends are doing it too! In fact, your friends brother is already working at a bank and can recommend you (he gets a small bonus for that). You slowly start to FOMO.

FOMO as a value indicator

Whats worse than missing an opportunity? Missing an opportunity while others around you don’t. Neighbours of lottery winners are more likely to go bankrupt. FOMO is real. Wealth and value are social characteristics. At the end of the day, things are worth what people are willing to pay for it. Bitcoin critics point out that bitcoin (and other cryptocurrencies) have no intrinsic value. Whats the intrinsic value of the US dollar? Some time ago, paper money could be exchanged in a bank for gold or silver. Not anymore. The paper money in your wallet, the bytes representing your money in your bank account, they have only social value. Its what other people will exchange for them that makes them valuable. Value is ambiguous, often determined socially, hence if others consider something valuable chances are it is valuable. There is increased competition for valuable things. You can also increase the value of things by increasing the perception of value/ scarcity. Every second rate salesman or marketer knows this. “Buy this! There’s only one left!”

At the career events I heard that I might be at a leading university, but getting a job at J.P. Morgan, McKinsey & Goldman Sachs isn’t as easy as getting into good university (which is of course only easy in hindsight). It is quite competitive. You have to read up on these companies and what they do. You have to pass tests. Practice interviews. All your friends are doing it, including that annoying idiot that always gets the best grades and does University level sports and just does everything perfectly and nothing wrong. You push yourself and have dozens of interviews in London (or some other big city) and write many cover letters and get advice from people on what to do. Its hard to get, therefore it must be valuable. The value of the job creates competition and the competition adds value to the job, but it also makes life difficult for recruiters. How do HR people select between dozens of detail oriented young adults from top universities who are dazzled by the promises of money and prestige? The only reliable way to get in is to expend significant effort in a recognisable way, that is to engage in Social Proof of Work.

What is Social Proof of Work?

First of all, Proof of Work is a key protocol in Bitcoin (and many other cryptocurrencies) and its game theoretically grounded implementation is one of Satoshi Nakamoto’s genius ideas. A simple explanation of Proof of Work in Bitcoin is that you have a computer guess numbers until you get the cryptographically correct number. Mathematically it is easy to check that the number is right, but it is hard to come up with the right number. That is the key element of Proof of Work: it is easy to verify but hard to do, hence a single number shows evidence of the resources you’ve invested. Today there are millions of computers humming along, spending more electricity than small countries, mining bitcoin, all because of the demands of Proof of Work. According to the Bitcoin Standard it is Proof of Work that makes Bitcoin a good store of value. All the effort and competition in mining for bitcoin, coupled with its innate scarcity, adds value to it. Yet the idea behind Proof of Work was not invented by Satoshi Nakamoto. In fact, the basic concept has been understood since times immemorial by parents, bosses, teachers, University recruitment specialists and authorities everywhere, and has been used solve a problem much older than digital cash.

Definition: Social Proof of Work protocols are social signalling mechanisms whereby humans can easily verify that another human has invested significant effort into a task.

Social Proof of Work is often expected in situations where:

Examples of social activities that can be Social Proof of Work:

Social Proof of Work can be divided into two types: useful Proof of Work and useless Proof of Work. Useful proof of work is work that has a function other than to prove the invested effort. In the cryptocurrency space useful Proof of Work is somewhat of a holy grail: imagine if all the computers guessing numbers for bitcoin were searching for life signs from far away galaxies or simulating complex biochemical processes. The problem with useful Proof of Work is that the goals of easy verification and usefulness often oppose each other. That is not to say that useful Social Proof of Work does not exist. For instance, memorising multiplication tables can be an example of useful Social Proof of Work. Even useless tasks may be useful because of their difficulty: it may be useful to do them for no other reason than that it improves physical and mental endurance. This is why the stereotypical Hollywood military recruits at bootcamp carry logs on their heads. Usefulness isn’t binary and often depends less on the task than on the individual doing it. A key question workers should be asking is, who is this expended effort useful for, and how?

You can also divide Proof of Work into relative and absolute Proof of Work. Do you want to prove you’ve put an absolute amount of work in or an amount relative to others? In cryptocurrencies you have relative rewards although there is an unambigous absolute measure of work (hashpower). In Social Proof of Work situations you can have both relative (if your exam is graded on a bell curve) or absolute Proof of Work (you pass the legal requirements for a drivers license).

Competition is a critical to relative Proof of Work tasks. You want to prove not that you want something, but that you want it the most out of all the other people that want it. Hence, as an investment banking analyst you’ll have to sacrifice sleep and health to prepare a powerpoint that does not really matter, to prove you want it the most. Not only that, you’ll also have to go to a top university for a job that can and used to be done by high school graduates. Having done the two years as an analyst at a grueling company with 100 hour weeks is the white collar equivalent of getting a gang tattoo on your face. It is a stupid sacrifice, but it shows commitment in an easily verifiable way. Moreover, competition also signals perceived value.

Why use Social Proof of Work?

The problem is ubiquitous: as an individual you want to follow an optimal strategy, i.e. a strategy that maximises your chances of achieving your goals, be it becoming a billionaire, getting a Nobel Prize, teaching a classroom of children or completeing a marathon. The problem is that the relationship between the optimal strategy and the observable results is not clear in the real world. If you get results (i.e. achieving goals or sub-goals), is it because you are doing the right things, or is it because you are getting lucky? Conversely, if you aren’t getting results, is it because you aren’t doing the right thing or are you simply unlucky? This is a complex problem in itself, but there is an added layer of complexity when you have to evaluate the work of others, mostly because know know less about what others do than what you do yourself. Consider that many would agree with the following heuristic: in a large company 80% of the useful work is done by 20% of the employees. Why don’t corporations, profit hungry and happy to cut costs, just get rid of 80% of employees (the useless ones)? Perhaps because the decisionmakers evaluating employees are simply lazy, but perhaps because it hard to tell useful employees apart from useless employees.

Why? We’ve mentioned that observed results are merely correlated with the amount of work put in. In some cases the relationship is linear: the more time you spend chopping wood, the more wood you will have chopped. In other cases, particularly with rare but highly profitable rewards, the relationship is far less linear. For example, pharmaceutical companies spend a fortune on research and development, but it is impossible to predict when a profitable drug will be found. Compared to chopping wood, research and development has a long iteration cycle. In areas with long iteration cycles, evaluating results within a short timeframe is foolish. As a thought experiment, lets say you require surgery to remove a tumour. You have a choice between two surgeons, Abdullah (successful 60% of the time) and Billy (successful 50% of the time). Lets say you have to settle for one, but you do not know which one is best. If you observe both Abdullah and Billy perform 1000 surgeries, it will be fairly clear that Abdullah is better. However, if you can only observe a two surgeries every two years, you’ll die before you can be certain which surgeon is best. Keep in mind, this is a simple example with constant success percentages, but in the real world it is entirely possible that the success percentages change over time (Billy gets better, Abdullah’s hands start shaking due to alcoholism) or circumstances (Abdullah may be operating on the easy patients with shoulder tumours, while Billy might be operating on harder brain tumours). In sum, it is difficult to judge performance merely by observed results.

Another problem is that observed results are often ambiguous. In the previous example we have an unambiguous result (patient died/ didn’t die). Most often however we do not have such clear binary outcomes. How do you evaluate the performance of a teacher, a UI designer or a compliance lawyer? Managerial culture has tried to come up with an objective way of doing so. They are called Key Performance Indicators. Anyone that’s worked with them know what the problem is. In the book the Tyranny of metrics Jerry Muller argues that as soon as a metric becomes a performance target perverse incentives arise because workers are encouraged to hack clearly defined metrics rather than to create ambiguously quantifiable value. The entire industry of Search Engine Optimising is just about gaming Google’s metrics. If surgery success rate becomes the surgeons Key Perfomance Indicator, Billy will stop doing brain surgery and will start working on liposuction. Models are simplifications out of necesity, and reality is often too complex for decisions to be made based on a simple spreadsheet model.

Most people understand that models are simplifications, and that KPI’s are models. Excessive focus on the easily quantifiable KPI’s results in performance that look great on spreadsheets and powerpoints only. If you spend too much time gaming useless metrics reality will shake you up. Or you might end up playing the grownup equivalent of the musical chairs game, the Bob Rubin Trade where you pass an ever increasing risk on to the next person(s) in the hope that you’re the one that ends up with the nice bonus and not the one that has the whole thing explode in their face.

Other than helping deal with ambiguity, Social Proof of Work is useful because past behaviour is a good predictor of future behaviour. If an individual has a history of investing a lot of effort on a useless task (say they’ve written an 80 thousand word thesis on Norse poetry in the 11th century) you can assume they will invest a lot of effort in whatever it is that you want them to do, even if you cannot measure it well.

For these reasons sometimes it is more convenient to evaluate effort instead of results. So how does one evaluate effort? People in positions of authority want to make sure people working for them work as much as they can, but often it is difficult to actually check up on it. From a teachers point of view, sometimes it is easier to test whether you studied than to test your unobservable knowledge. From a recruiters point of view, does an applicant actually want to work for the Equity Research desk at Credit Suisse, or did she just change a few words in the cover letter she sent to McKinsey? You could spend hours with your subordinates, observing their work, questioning your students, calling your recruitment candidates to ask questions. But there is a limited amount of time. You want to be able to check effort, but you want to be able to do so easily. Recruiters often talk about “cultural fit”. This can mean “is the person pleasant enough to work with”, but it also means: “does he/she talk the talk”. It is important for consultants and bankers to talk like consultants and bankers. It is evidence that they’ve invested effort in finding out how consultants and bankers talk. It is Social Proof of Work.

Hacking Social Proof of Work

In any profession, 90% of people are clueless but work by situational imitation, narrow mimicry & semi-conscious role-playing […] Nassim Taleb

In practice most evaluations of human performance are a combination of looking at heuristics for both observable results and Social Proof of Work. They are both imperfect measures, but together they do a relatively good job at separating performers from non-performers. Nonetheless, unlike Bitcoin’s Proof of Work protocol, Social Proof of Work can be hacked and often is, both by people whose performance is evaluated and by those who are ostensibly evaluating performance. Moreover, it also creates some paradoxical situations which are familiar to anyone that has worked in a large organisation, such as that the more absolute amounts of work get done, the lesser the percentage of useful work.

Why is there so much useless work in competitive environments? In a competitive relative Social Proof of Work situation (e.g. doing a 6 month long internship at a Prestigeous Institution or a postdoc position at a Prestigious University) there is an ever increasing amount of effort that needs to be expended in order to stay ahead of the competition. However, often there is a limited amount of useful work to do. Of course, an enterprising individual can always find more useful work to do, but the usefulness of the work overall asymptotically decreases. Hence the increased amount of work done due to competition reduces the proportion of usefully expended effort. The more competition, the more useless work. There is empirical evidence for this, in the highly competitive field of academia Ioannides estimates over half the published scientific are false. False findings aren’t necessarily evidence of useless work, but their predominance is. This is an example of individuals hacking the system, evidenced by the fact that people that do nothing useful whatsover can thrive in this environment.

But not only employees can hack the Social Proof of Work protocol. So can employers. Competitive relative Social Proof of Work tasks allow employers to lower their employment costs by creating social wealth and making probabilistic payment possible.

How does one create social wealth? Success in social competitions leads to prestige, which is another way of saying social wealth. Every year there is intense competition to be admitted to university to become a medical doctor in Hungary, even though doctors are disgustingly underpaid. You can get away with paying your employees less if you increase the social status of the job you are offering. The best example of this are the “feel good” moralising human rights/climate change NGOs that can literally get away with not paying their entry level employees (interns) because of the social prestige and intense competition attached to the job.

Probabilistic payment is another way in which you can reduce the cost of your employees. Paying employees probabilistically can be done by creating artificial scarcity (e.g. hire 5 interns to audition for 1 job opening or train 25 PhD’s for 1 professorship). The way it works is you offer your employees the chance of a large reward, be it monetary (e.g. “you might make it to partner at our firm!”) or social (e.g. “you might become a tenured professor!”) or some combination of the two (e.g. “ you’ll live like a king on your salary as a development expert in this undeveloped country”). Of course, you then only give a small percentage of your employees the rewarding role, but only after years of highly competitive, underpaid (and increasingly useless) work. In fact, if you are cheeky enough you don’t even have to give any of the employees the rewarding role, a friend of mine worked 100 hour weeks for six months as an underpaid intern only to be told that “although you were the best intern, there are currently no openings”. Probabilistic payment is dangerous to employees because it alters the rules of the game: your employer is no longer trying to select the best performing individual through imperfect measures, but rather is running a lottery where individuals pay with their work. This means that there is less of an incentive to reward the hardest working individuals.

Life in times of Social Proof of Work

Social Proof of Work arises out of situations with ambiguity where visibly expended effort is used as a proxy to results. Ambiguity is inevitable, we live in an uncertain, complex world and there is nothing we can do about it. In fact, the more complex our world becomes, the more ambiguous the relationship between work and results become. The results of hunting and foraging are less ambiguous than the results of a credit analyst. Hence, Social Proof of Work is probably inavoidable as well, and that is not necessarily a bad thing. It is useful to be able to show that one has spent effort without necessarily having to show results, precisely because we live in an ambiguous world where the relationship between effort and results is opaque.

This opaqueness can lead to all sorts of problematic situations. One is when there are no consequences for consistent failures to produce results, such as in academia, where many fields have stopped producing value and are engaged in situational imitation, narrow mimicry & semi-conscious role-playing, (a.k.a. a cargo cult) in other words, in a vicious cycle of useless Social Proof of Work (again, see this. This is clearly unsustainable. Peter Thiel has long held that the education sector in the US is a bubble. In terms of outcomes, a bubble is not much different from a Ponzi or a pyramid scheme. The last ones in are left bagholding.

Why are we attracted to pyramid schemes? It is not just greed, in fact I would argue it is mostly ambiguity. Humans do not like ambiguity, and crave structure. Because we live in a complex world, whenever a clearly defined structure exists it exists because it has been structured. And if it has been structured, it has been structured by someone, and whoever structured can’t help but to structure it in their best interest. Structure salesmen include prophets, university recruiters, graduate recruiters, ideologists, academics and Team JUST. Some structures are fairly innocuous: driving on a particular side of the road makes travelling safer. But other structures clearly benefit some more than others, such as the vision that involves moving to London after university, working long hours, getting paid probabilistically, paying 30% income tax to pay the pension of your landlord, commuting long hours, paying 30% of whats left after tax on rent to someone that got into the property ponzi before you did.

In a complex world ambiguity is unavoidable. As long as your life has a social component to it, Social Proof of Work is probably unavoidable too. But it is important to be aware of it. If you are not, others will be and you’ll end up making the career ending mistake of working hard instead of visibly working hard in an industry that favours the latter. Structure is desirable and can help overcome difficulties with ambiguity. Yet it is best to create your own structure, or at the very least to be aware of what the structure is and who constructed it. Team JUST and graduate recruiters do not have your best interests at heart.

UPDATE: The first round of FOMO3D has finished.

An enterprising person (or team) has won the first round of FOMO3D, winning 10500 ETH (approx 3m USD). How did they win FOMO3D? Apparently they exploited a few properties of the Ethereum blockchain and spammed it so no other transaction could get through. Some games can only be won if you game the metrics, cheat, or you set up your own game.

comments powered by Disqus