Posted on by admin

Art Isn’t Magic

There’s a real problem with the way we talk about creativity and art, especially when it comes from people who have never tried to make it themselves. Art of every variety is frequently presented as a borderline mystical activity, as a shamanistic ritual to extract some fundamentally unknowable substance from some other world and merely capture it here on ours. Artists are treated as a different type of person entirely, like those born of special blood in a fantasy story, possessing an innately different ability to perceive reality.

Make no mistake, this is nonsense of the highest order. Art is not magic, it only appears so to people who have made no real attempt to understand it. It is not a figurative comparison to say that people who think art is magic are like Stone Age hunters seeing an airplane and assuming it’s a mystical being, because what they’re experiencing is literally the same mental process. They have mistaken a property of themselves (not knowing a thing) for a property of the art (an unknowable thing) and, in doing so, convinced themselves that because they don’t understand the art (or the process behind it), that is must not be comprehensible at all.

When we feel that we lack the faculties to comprehend something, we tend to dress it up in myth and mystery, to present it as not being something that we don’t understand, but rather something that no one can understand. Sometimes it’s merely that only those people can understand, and then we conjecture wildly about what delineates that category. School teachers do this all the time with their utterly asinine attempts to categorize students into left vs right brain, the thoroughly debunked “learning types”, or meaningless groups like “math people”. The implication is that the failure cannot be on the teacher or curriculum, but instead must be a fundamental type error, like trying to put a keyboard into a CD drive.

This has tangible effects on young people, and I was one of them. Despite my love of storytelling, there was a time when I became convinced that I was too “logical” to be a writer. When viewed in retrospect, this is obviously absurd, but it really gripped me at the time. Artists were these floating, spiritual beings, pulled about by the muses, touched by magic, ever dreaming. Art, it seemed, was a thing to be felt, a wringing of pathos onto page or canvas, a vomiting of emotion and energy. Like a prophet hearing the voice of a god; if you aren’t certain that he’s talking to you, then you must not be one of the chosen ones.

Art is not magical, nor is it a gift merely bestowed on some and not on others. “Ah, but Kyp,” you might find yourself saying, “nobody really thinks it’s magical, it’s just described that way to capture how it feels!” But when I say that art isn’t magical, I don’t just mean that it’s not a supernatural power, I mean that it isn’t magical in the metaphorical sense either. The problem with thinking of it as somehow special is that it makes it seem like something someone else can do, but fundamentally not something you can do. Once you’ve primed yourself to believe that “artist” is a type of person, you’ve separated it out from the mundane and that comes with certain implications of how it will work, what rules it follows.

Consider this: Would you ever use the term “electrician” to describe an entire group of people, whether or not those people actively work in the field at all? When someone asks why you decided to not pursue your lifelong dream of studying to become an electrician, would you answer “well, I’m not an electrician”?

Notice how circular that sounds when you swap in a “mundane” profession? You didn’t study to become an electrician because you’re not… an electrician. Well, of course you’re not an electrician if you never attempted to become one, that’s just how causality works. One is obviously not born an electrician, it’s a thing you become through gaining knowledge and developing skills just like everything else.

A pet peeve of many visual artists and musicians that I’ve met is the phrase “oh, I wish I could do what you do, but I can’t.” The statement “I can’t play the trombone” is not a statement about you fundamentally, it’s a statement about you right now. It’s a description of your current knowledge and your current skills, a collection of facts about your present state. You are making an inference that is both wildly incorrect and a bit insulting, namely that the artist in question has always been that good, rather than getting there through years of study and practice.

This is one of those beliefs that I think people can only defend if they don’t really acknowledge what they’re defending. If you simply say, “some people are just natural musicians”, it gets passively absorbed by people as “one of those things about the world that is true”. But break that statement down into what it actually means, and what are you left with? What does “musicians” mean in this context? Surely it means “someone who has the skills and knowledge to play an instrument (or sing, or compose) well”. Are you genuinely proposing that they emerged from the womb already knowing how to play the piano? Or is it more likely that what you’ve mistaken for innate ability is really just that they have learned it more easily than you believe you would have?

“Well, come on, that’s what we’re actually saying, isn’t it?”

I don’t think it is, really. When pressed, perhaps people will walk back their belief to “some people have an easier time becoming skilled at X”. But these same people are still making and justifying decisions as though what they really believe is that a box is checked when they’re born and they’re either an artist or they’re not. If it’s merely that they think that some people will learn faster than others, that would be so trivially true as to not be worth stating at all.

The reality is, for a truly average person, it’s possible to become quite good at almost anything given enough time and effort. The obvious exception is unalterable physical limitations in activities that are >90% physical ability. Even in things that require some physical component, while your lack of physical ability might prevent you from being in the top 100 people in the world, it almost certainly doesn’t prevent you from the top 1% as long as you are willing to work harder on skills and knowledge. You’re not obligated to pursue any particular skill, artistic or otherwise, but you can’t rely on the notion that you fundamentally are unable to do it. And that’s not even a bad thing. We all have limited time and energy and choosing to not commit to becoming great at a specific thing is okay. But ultimately, it’s a question of what you want, not of what you’re capable of.

Resist the urge you may have to reframe this issue in such a way that you can continue to say “no, I’m just fundamentally incapable of being good at that thing.” Instead ask yourself why you would even want that to be true. That you are able to become a very good painter with enough study and practice does not mean you are obligated to do so. It doesn’t even mean you should in some greater moral sense. It merely means that you cannot pretend that it was never your decision as a way to excuse yourself from the responsibility of making the choice.

What’s lost in our elevation of art to the magical level is that its beauty is derived from the fact that is real. The wonder of art is directly derived from the real world, not from some mysterious aether floating outside normal perception. This is true whether or not the subject of the art has ever existed, as I’ll expand on later. So, don’t get fooled by this idea that you are not able to create art. You may just not be able to yet.

Posted on by admin

The Parable of the Widget Certificate

Imagine a theoretical industry making theoretical products called widgets, and suppose as well that they are looking for people to employ for entry-level widget-maker positions. The primary skill of these widget-makers is testable, but not easily testable, and is difficult to fully establish from just a standard resume-and-cover-letter style job application. It is possible to figure out if a prospective candidate possesses the qualities that are needed for the job through testing and interviews, but it’s relatively time-consuming and expensive.

So the companies in the industry do this extra testing, but they aren’t happy about it, and it limits the number of employees they can consider. And there are more competent young widget-makers than there are available positions for them, so it’s useful to distinguish between a good one and a great one, but it’s also crucial to test to make sure that you’re not hiring an unqualified one.

Note that this process is essentially only needed for new widget-makers, as senior level positions are judged on the quality of work done at a lower-level. And assume as well that the industry is expanding, but at a rate exactly (or very nearly) offset by retention, such that in the industry as a whole, the number of new widget-makers needed each year is the same. Which is to say that however large the industry gets, it still only hires an absolute value of, say, 100 new widget-makers each year.

Then an enterprising group comes along and says “It costs you all this time and money to make sure that someone is qualified for a position, and you suffer from only having staff to do it part of the time, or having generalized recruiters that don’t know how to determine the best widget-makers. Instead, let us first formally certify widget-makers as possessing the knowledge and skills that they need, and we’ll be able to do it efficiently because it will be our sole focus.”

This is an appealing pitch if you’re Widgets, Inc. because it simplifies the process for you, means you don’t have to develop your own tests that you will only use a few times a year, and reduces your costs. But if you’re the new Widget Certificates provider, where does your profit come from? You could charge the businesses for the service, but they’re the ones that you have to convince first, so you want it be as enticing to them as possible.

So why not charge prospective employees? This makes sense since you’re screening them for the knowledge and skills they need to make widgets, and once they have your certificate, they’ll have a much easier time getting a job than if they had to convince an employer to give them a chance to prove it. Because, after all, your certificate expands the pool of employees a widget company can now consider by simplifying the applicant review process. And if you have a certificate proving your knowledge and skills and no other applicants do, you have essentially already passed their internal test, giving you a specific advantage over the competition.

So a benefit is provided for the employers by reducing costs and an advantage is created for employees at the cost of paying for certification. ‘Benefit’ and ‘advantage’, it should be noted, are distinct concepts here; employers pay nothing for their benefit, but employees pay something in return for an advantage. What the certification provides for an employee is, essentially, a service, a transaction, or an exchange of value. What it provides for the employers is a positive externality, a benefit they are receiving outside of the closed transaction between Widget Certificates and prospective employees.

With a distinct return for both employees and employers, this certification system naturally expands, with more certification providers popping up to fill the space as more people begin to use it. Very quickly, widget companies only want to review employees that have earned it. A clear line of distinction is drawn between qualified and unqualified widget-makers, defined by who can and cannot successfully complete their certification. This solves one of the major problems of the widget companies, which is having to test to be sure that whoever you are hiring is at least not an outright bad widget-maker. The certification is the test now, and having or not having the certification becomes the equivalent of passing or failing the test.

We now have three major groups of widget-makers, each in a different situation.

First, you have the bulk of your widget-makers who passed the certification (2nd-level). They are competent and capable, but not exceptional. The initial creation of the certificate helped them by removing the unqualified widget-makers from consideration and thus decreasing the total pool and making them more likely to be hired.

But there are still more potential 1st-level and 2nd-level employees than positions, so this certification process has settled in on a new normal: it is necessary to have a job, but not sufficient on its own to be guaranteed one. Some 2nd-level employees with certifications will not be hired. However, because the companies are no longer running their own tests to separate the top from the middle, the median widget-maker can now occasionally get a job over a great widget-maker.

Second, you have the great widget-makers (1st-level). The creation of the certificate didn’t help them, because under the old system, they were already extremely likely to get a job. And when the companies were testing internally to weed out the bad (3rd-level) widget-makers, they also gained insight into who was a great widget-maker versus a merely good one. Before the certification, a great widget-maker was virtually guaranteed a lucrative position, but afterwards is not significantly distinguished from the rest of the pool.

Third, you have the potential widget-makers who did not pass the certification (3rd-level). Their situation has gotten dramatically worse, because while the previous internal company tests frequently locked them out, some of them could still get positions at widget companies with less strict guidelines if they were a good fit for the company itself. Now, the certification process has set a minimum bar for employment in this lucrative industry, and being even slightly below it means the difference between financial success and poverty.

The 1st-level widget-makers are dissatisfied with the new equilibrium, which makes sense; the new system doesn’t benefit them and they feel that being better at the job should translate into something tangible. Seeing this opportunity, several of the certification providers decide to break off, make their certification test much tougher, and charge a premium for it. The increased cost for a higher certification is worthwhile if a widget-maker is good enough because separating oneself from the bulk of certified applicants will ultimately result in a larger return long term.

At the same time, the 3rd-level widget-makers, having now been shut out entirely, seek new recourse to regain their lost opportunities. But there isn’t a financial incentive for any certification provider to break off and provide a new certification that they can pass, because no one would sign up for it; if there are already more employees than needed at the current certification level, no company would ever hire one from a recognizably lower level, and thus no one would pay for the certificate in the first place.

What the newly unqualified really want is to mix back in with the main pool, to become indistinguishable from a 2nd-level widget-maker so they can go back to having a job. Naturally, the 2nd-level widget-makers don’t want this to happen, because they’ll then have an even harder time getting a job. And the widget companies don’t particularly want this to happen either, as they’re no longer doing the internal tests to make sure they don’t hire an unqualified employee, so there is no failsafe if the certification system isn’t accurately signaling knowledge and skills.

But the certificate providers aren’t financially tied to the companies. Remember, the companies benefit from an externality of the certification system, but they don’t directly contribute to it. Their recourse if it doesn’t work is to stop using it and go back to their internal testing. But the internal testing was expensive and time-consuming and requiring certificates is easy and free; for it to be worth switching back, the value they’re losing has to be greater than the cost of doing internal testing, and that lost value is going to mostly be intangible and hard to detect. If you assume even a generous rate of every 2 dollars of intangible value lost is as noticeable for a company as 1 dollar of tangible value, it would still take a very large amount before they’ll consider reverting to the old system.

Well, what about the 2nd-level widget-makers? If the certification providers start reducing the test difficulty subtly to make more 3rd-levels blend in with them, why wouldn’t they just stop paying for the certification since it wouldn’t really be helping them anymore? Because it would still be helping them to not be shut out of the system entirely, like all of the 3rd-levels currently are.

The introduction of the certification turned the nuanced gradient of employability into a hard cutoff line. And the companies are going to keep requiring it even after it no longer provides the same value because it doesn’t cost them anything. Perhaps all of the 2nd-levels could walk away from the system en masse, but no individual can, because then the system will continue and they’ll be shut out of it.

Naturally, the certificate providers begin lowering their standards to increase their sales and the already over-saturated employment market becomes even more inundated and uniform. The companies aren’t thrilled, but are glad that at least some of the 3rd-levels are still able to be filtered out. The 2nd-levels see the writing on the wall though, and look to the 1st-levels solution; while they cannot pass the 1st-level certification, they can push for a middle ground. Soon enough, a 1.5-level pops up, above the current 2nd-level but below the 1st-level. The process of stratification accelerates.

Let’s fast forward a bit to the inevitable result. Because each group stands to benefit from having the highest level certification they can get, the tiers continue to accumulate until eventually there are, for sake of argument, approximately 20 tiers of certification. This is not to say that there are only 20 certification providers, as there are thousands, but that there are 20 certified levels of widget-making quality. At this stage, a couple of things happen.

First, the cutoff bar has become increasingly precise, such that widget-makers who would previously have been median, 2nd-level candidates and thus able to be employed might now be shut out entirely. Remember, there are more prospective widget-makers than there are jobs, and the increasing stratification of these candidates pushes the line closer and closer to the point where exactly as many people get certified as there are jobs. This would seem, on the surface, to be a positive change. After all, isn’t it ideal that no one attempt to get a certificate if there won’t be a job waiting for them afterwards? But there’s a catch, which we’ll get to.

Second, because of the stratification of certificate tiers, it is increasingly difficult to tell them apart. In the same way that someone with even middle-level knowledge of basketball can quickly differentiate between an NBA star, a mid-level college player, and a person playing in their driveway, anyone with a reasonable understanding of the widget industry could easily tell the 1st, 2nd, and 3rd levels apart. There was blurriness at the margin of course, and a 2nd-level that was almost an 1st-level (or vice versa) might have been significantly affected by the rigidity of those tiers, but the bulk of people were classified correctly.

Now, suppose that the number of, for example, 10th-level widget-makers is greater than the number of available positions, but the number of 9th-levels is less than the necessary amount. Logically, 10th-level is then the cutoff, with some but not all 10th-levels finding jobs and no one from 11th-level or lower. This would be a simple decision for a company, but how do you know if an applicant is in the 11th-level or the 10th one?

At this point, let’s clear up an assumption that these levels are based on a strict criteria doled out by some governing body. Make no mistake, the levels are “real” in so much as the semantic purpose of them is served by distinguishing them; two people discussing a 10th-level skill at widget-making are probably talking about roughly the same thing. But it’s very hard to easily know what level someone is actually at, just as most skills are hard to precisely evaluate quickly. A quick look can get an approximation, but to increase confidence levels, more time and effort is needed.

So there’s nothing to stop an 11th-level certification program from saying that they provide a 9th-level certification, because the levels aren’t official, they’re just an idea formed around a particular amount of knowledge and skill. And there’s no way to enforce any kind of official standard either, because the material is still too subjective for that sort of rigidity.

But surely people would realize that the certification providers are lying, right? After all, if they’re really testing at the 11th-level and are claiming to be providing a 9th-level service, then just compare them to 9th-level certificates and you’ll see the difference. Except that each one has a different test, not merely a different score requirement. They feel there are many different approaches to making sure widget-makers know their stuff, and refuse to all conform to a single test.

But in fact, many are not merely claiming to be slightly better than they actually are, but many levels higher. Without any meta-level standards, there’s no simple way to know which certification providers are being honest, and there are too many of them for any hiring manager to keep track. Companies can’t afford to risk hiring someone whose certification says they’re a 5th-level, but is really a 12th-level.

So another enterprising group decide to rank the certification providers, so that companies will know how good or bad each certification is without having to rely on the marketing of the certification providers themselves. But while this idea catches on, it never really solves the problem. Ultimately, subjectivity has to enter into the rankings, and the list shifts around a lot from one year to the next.

While companies really like having the list, the certification providers are heavily rewarded or punished for their placement on it. This strongly incentivizes them to make changes to appeal to the ratings group, because the companies are now looking to that group to decide who is talented and who is not; being perceived as a 3rd-level program instead of a 7th-level program significantly increases what said program can charge.

The ratings group is genuinely trying to provide a useful service, but what the certification providers do is complex enough that no simplified rating system can capture the full scope of it. And the ratings have to be simple and easy to parse because that’s the whole reason the companies want them. But this necessarily leads to two negative outcomes.

First, the certification providers are strongly incentivized to only make improvements in the areas that the rating group cares about, even though those might not be the areas that correlate with quality. In fact, they are instead frequently just the metrics that are most convenient to measure. This allows the certification providers to easily convert money into a higher rating and for ratings groups to minimize their work load, but it doesn’t provide a tangible benefit to either the companies or the applicants seeking certification.

Second, because these ratings are public and well-known, the applicants are incentivized to use them as the basis for deciding which certifications to try to pursue. But this creates a feedback loop, since the perceived demand for the certification program is part of how they are scored for the ranking system. This means that once a certification provider gets a high rating, it will also get increased demand, allowing it to maintain that high rating indefinitely off of that original momentum.

Let’s fast-forward again. Because of the way this system is set up, and the strong push each certification provider has to give the best-sounding certificate they can to whoever pays them, things slide increasingly towards meaninglessness. It becomes too difficult to track the quality of certified widget-makers, and the correlation between the certification and the quality of employee trends towards zero. Companies can no longer really trust the certificates as a sole method for hiring, so they decide to return to their internal testing.

But they don’t get rid of the certificate requirement. Candidates will now have to pass the internal testing to be hired, just like they would have originally, but now they have to have a certificate to even make it to that level of internal test. At this stage, the certificate is no longer about picking the right candidate, but rather filtering out as many as possible to reduce the cost of doing the internal tests, which are the actual decider.

And the situation is worse, because in the time that has passed since the system began, there have been court rulings that decided that tests could only be used if they could be conclusively proven to be very strictly and carefully catered to exactly what the job requires. The goal is to prevent prejudices from being worked into these tests, which would impact certain groups at a higher rate. The hiring departments of the companies shrug and say “No big deal”, because of course the tests are catered for the job.

But the legal department says no. If someone challenges the validity of the test, they contend, it won’t really matter if the company will win the resulting federal lawsuit, because the cost of litigating it will be enormous. And even if the test is not fundamentally biased, if the outcome of it appears to show a bias, they’re still liable. So the old internal tests are out the window.

What now? The original tests are gone, and the certification system that had replaced it is no longer doing anything useful to filter who is and isn’t qualified for a job. There’s as many candidates as ever to sort through, but the hiring departments have gotten used to having convenient filtering mechanisms, and now they have nothing. They won’t get rid of the certification requirement, that much is certain; it still costs the company nothing and if it filters out even 10% of the applicants, it makes their jobs that much easier.

What everyone agrees on is that the certificates were worth something at one point, even if they’re not worth much anymore. And given enough time, a decent amount of the applicants ended up being pretty good at their jobs. So a new requirement emerges to help the companies filter down to those who either passed the certification before it became ineffective or those who have already proven their ability in the industry: work experience. The best companies can ask for more, the worst for less, but each of them adds a new requirement of no less than two years of experience working as a widget-maker. They dust off their hands, satisfied, and the last person left with a dream of widget-making is locked out of the industry for good.

II.

The comparison here is not 1-to-1, as there are obviously differences in how our current world got to where it is. The real situation is much more complex. But the focus here is on the different motivations and intentions at play and how they interact with each other to create unintended consequences. Every group in this story is pursuing a signal, but in the process, they are getting horribly bogged down in noise, and they continue to add increasing layers of proxy and abstraction until the signal they’re detecting may not even be there at all.

Let’s start with motivations: as is usually the case, all the actors here have reasonable and non-destructive motivations. The companies want to hire the best employees possible without spending tons of money doing it. The applicants want a way to prove that they are able to do their job so that the companies will be more likely to hire them. And the certification programs want to provide this service and profit by doing so.None of these parties is pursuing something immoral at the baseline, and there is an opportunity for mutual benefit, for a system that helps all of them at the same time.

The reason the whole thing collapses, however, is because of perverse incentives, unintended consequences of well-meaning actions caused by a poor alignment of goals and methods. The arrangement creates a perverse incentive right from the start, because the certification programs are never financially tied to the companies. If they were, their most important goal would be to ensure the companies were satisfied with the employees that they were receiving, lest they stop paying for the service. But since the incentive is instead to keep the applicants happy since they’re the ones paying, the service being provided is not “designate who is best qualified” but rather “provide the best route to qualification.” If the system was bad from the very start, this would fall apart immediately because the applicants would never pay for the service, but because it falls apart slowly, the system becomes entrenched.

The fact that the certificates cost the company nothing creates the perverse incentive to continue using them until they are specifically costing the company more than what it would cost to do their own tests. Even though the certificate cost balloons dramatically for the applicants, it stays locked at zero for the companies. And the fact that getting lower quality employees is a real cost only partially factors in, because the specific cost of a bad employee is hard to quantify. This is compounded by the fact that the part of the company making hiring decisions (and thus decisions about the cost of running a test) is frequently not the part of the company that ultimately winds up paying the cost for a bad employee, making it even harder to tie together cause and effect.

What the company is after is a signal of worker ability, and that turns out to be a somewhat difficult signal to pick out, so they turn to a proxy. But such a proxy is, by it’s nature, an approximation of the underlying thing and should never be mistaken for the thing itself. It’s just a way to take a difficult concept and reduce it down to something easier to process. The certification program provides that, but it starts them down a dangerous path of trusting a proxy without having a real way to make sure that it is still representing the thing they’re actually looking for.

To borrow an example of this idea, think about the gas gauge in your car. You have an indirect relationship with the gas tank in that you don’t have a way to directly perceive it while you’re actually driving the car. The gauge provides a proxy, a simple visual representation of the underlying truth of how much gas is in the tank, but it is not the gas tank itself. We come to rely on it and we don’t really make that distinction anymore, which is reasonable. But if the gauge breaks and we continue to trust it as though it isn’t even possible for it to break, the only way we’re going to realize that it’s not accurately tracking the gas tank anymore is when we run out of gas unexpectedly.

When the system starts to break down, the companies make a crucial error: they turn to another proxy to represent their original one. Because there are differences in the programs (the underlying complex truth) that they need a simple approximation of (the proxy), they turn to the ranking system to tell them which programs are good and which ones aren’t. But adding additional approximation makes the signal harder to detect, not easier. It’s like adding an additional gauge to judge the quality of your gas gauge instead of re-evaluating why it broke in the first place. The further you get from the signal, the harder it is to be sure there’s a signal at all.

When they eventually realize that even the ranking system is not able to accurately track the signal anymore, they start to move away from it, but now they’re not able to return to their old method because the law has changed. So with the old system banned and the new system broken, they get desperate and start to look for the signal the only way they have left, by trusting that other companies have already found the signal for them, via work experience.

The error in this logic is that if the other companies have a good way to track the signal but you do not, then why not just use that method yourself? Perhaps they keep it a secret, but if they won’t reveal their method for tracking the signal, how can you even be sure that their system works at all? At best, you’re left to hope that they found the signal back when it was easier to track, and that they have removed the noise by firing anyone who couldn’t handle the job. But everyone’s life experience should make it clear that being a terrible employee is frequently not enough on its own to get people fired. So should “managed to not get fired for two years” really be considered a useful metric, much less the most useful metric?

The increasing distance from the signal also makes people more superstitious about where the signal is actually coming from, so they gather together a lot of these questionable sources as an attempt to make up for the fact that none of them are reliable. Obviously this does not work, because many inaccurate things added together does not make an accurate thing, but this behavior persists; humans are very good at sorting data into patterns and excluding the bits of evidence that don’t fit. Most employees hired under this system do wind up being okay at their jobs, so it looks, at a glance, like the system works. But what if the reality is that picking applicants at random would result in just as many good employees? The situation would look the same, and no one is monitoring for that possibility.The experiment has lost its control group.

So now imagine you’re a young and talented person ready to get into the widget-making career path. Every company requires that you spend a large sum of money to pass a widget-making certification, where the difference between a passing grade and an exemplary grade is ultimately not visible; either way, the certificate reads the same.

After you finish that process, you will then go forth into the industry only to discover that, in a classic Catch-22, every single company wants you to already have experience doing the job before they’ll hire you. Perhaps you should have gotten a better certificate, but it was four times as expensive and you’d have been paying it off for decades. You consider an internship, only to discover it’s harder to get than the job (because most businesses will only take current university students) and also doesn’t pay anything. So what exactly are you supposed to do?

III.

You may be wiping your brow and sighing in relief that you don’t work in this hypothetical widget industry. But the truth is, you probably do. With the very few exceptions of professions that either have a rigorous schooling process that puts out less students than there are jobs (medical school) or where your skill is largely technical and demonstrable (engineering), you work in this sort of widget industry. You may have started working in it before it reached its current stage or perhaps entered it through some rare or unique circumstance where you didn’t encounter it quite this way. But this is nonetheless the way it works for most industries.

Perhaps you think there’s a fatal flaw here; my example is just of a certification you have to pass, but the college system is there to actually teach you things. But this is extraneous information. Even if college was actually teaching students the skills that they need for the job, the colleges aren’t actually filtering these students enough so that their degrees prove to future employees that they have those skills.

And the evidence isn’t good that colleges are imparting much useful knowledge regardless; college dropouts make only a small fraction more than those who never attend and a great deal less than those who complete their degrees. If college were imparting knowledge that mattered, you’d expect every year spent there to be a meaningful increase to a student’s value, not a sudden spike only after the degree is reached.

But that’s exactly what you’d expect if college is just a way for company’s to offload the work of finding a signal about applicant quality. You get a small bump for getting into college at all, because that’s a signal of intelligence, and a bigger bump for completing it, which is a signal of your conscientiousness. But college isn’t transforming you into something, it’s merely filtering you down. This is why suggestions of universal college attendance are so misguided: if a filter is no longer effectively separating the thing it was designed to separate, the worst possible solution is to just suggest that nothing be filtered at all.

And that’s what has already happened in large part, because college has become so inaccurate as a way of filtering out talent and intelligence that companies won’t rely on it anymore. But because it’s free and easy for the company to request, they also won’t remove it from their requirements. It moves into the category of necessary, but not sufficient; you need it to get a job, but it won’t get you a job on its own. But remember, the college degree, from the perspective of the companies, was only ever a filter, not an ends unto itself. This is the equivalent of suspecting the filter doesn’t work and putting another filter on top of it without taking the old one off. It’s mistaking the tool for the end goal, as though it had some sort of intrinsic value beyond itss use as a filter.

From the perspective of the company, that’s a reasonable strategy, just in case the old filter is still doing something. But the US alone spends more than 650 billion dollars on post-secondary education annually. That’s the cost per year of keeping that old filter on that no one is really confident is still doing anything. But you have to pass through it anyway, or you can’t reach the second filter.

To talk about the second filter, the one that really counts, we have to talk about the worst system ever devised for determining employee talent: the resume. Your standard resume contains only two broad types of information.

The first is the intangible things, the skills that you have, the reason you’re well suited for the job, and the list of things you accomplished at your previous job. These are almost entirely worthless. It is very easy for people to be dishonest about their intangibles because it is very hard to prove that someone is objectively lying; if you say you have technical writing skills and you don’t, you can at least say you thought you had them, from your perspective. Because people can lie about these, it doesn’t really matter if you do or not, because no one really believes what you say. Or worse, they are unimpressed with it because you were honest and only listed your actual talents while everyone else created a long list of fictional talents.

The second part is the concrete information, which essentially reduces to your education and your work experience. This is your entire resume, because it’s the only part that the company can actually verify, which means it’s the only part that an applicant is actually discouraged from lying about. The rest of it is just decoration. If you don’t believe me, consider that for a large percentage of medium and large sized companies, a person never even reads your resume unless this information is considered good enough. There’s a reason so many job applications online have you fill out your education and work history in drop down menus in addition to including it on your resume: they use sorting algorithms to remove resumes that don’t include enough experience or education so that no one has to bother with them.

So just to make it to a non-zero chance of having your application read, you need both the education, which we’ve established is neither useful nor sufficient, and the work experience, which you cannot get because all of the entry-level positions require you to already have said experience. But what if you’re really, really good at it? What if you have so much talent that you’d be guaranteed the job if you could just get an interview, but you don’t have both the education (which still doesn’t do anything on its own) and the work experience? Surely there must be some way to circumvent the system, because there clearly are new employees being hired in these positions.

Without any other options, you decide to try to take a truly bottom-level job unrelated to your goal, no college or experience required, at a company that does what you want to do, with the hope that when the time comes, they might give you more of a chance because you already work for them and can showcase your work ethic and talent. And I’ll grant that I can’t say conclusively that this never works, because situations vary. But what if it doesn’t?

Well, the good and bad news is that there’s always nepotism. You can always get the job from someone you know, after all, and it’s not true that everyone hired for nepotistic reasons is incompetent. But that was never the problem with nepotism, because of course it doesn’t have a 100% failure rate. The problem with nepotism is that it heavily biases the in-group and that it fundamentally isn’t a path that can be open to everyone. Not everyone is going to know someone at the company already.

And supposed solutions related to networking are misguided at best, because the fact that networking is technically open to everyone doesn’t mean it practically is. While trying to exceed 10,000 connections on LinkedIn could result in an opportunity, that’s still heavily gated by the biases and lack of openness of the people you’re trying to connect with. Some people who would be very good at a job may not be very good at going to industry cocktail parties and schmoozing up to hiring managers. Most jobs don’t really need you to be a socialite. And no one really thinks that should be how people get jobs, or we wouldn’t have the application system at all. So why do we pretend like that means the application system isn’t horribly broken?

Networking isn’t a safety valve that makes the hiring system work when it would otherwise fail. “Just network” is a convenient excuse for people who don’t want to acknowledge that the system in place is fully locked in a failure state. If the only way for you to get an entry-level job is to befriend someone who is in a position of power, then you’re totally at the mercy of that person, forced to conform to what they want personally rather than what is important professionally. That rewards people who can play the totally tangential game of networking rather than people who have talent for the job itself. And do we really need to add even more rewards for that sort of chameleon behavior, where pretending to be what your boss likes personally is the best way to move up? Does that need to rule the process of getting the job as well?

If you’ll permit me one more analogy, imagine that you’re a world class quarterback trying to make it as a pro. But before the team will give you a tryout, you have to go through their quarterback program with their scouts. And at the end of the program, you have to throw the ball as far as you can, which in your case is an impressive 70 yards with a perfect spiral. But the scouting report doesn’t say how far you threw the ball, it just says you threw it more than 30 yards. So then, when you get to the stadium, the manager won’t let you tryout. Sorry, he says, but you need to throw way farther than 30 yards to make the team. Nevermind the stupid scouting report, you say, watch this, and you throw a beautiful 70 yard pass down the field. The manager watches it, expressionless, shrugs, and taps the clipboard in his hand. Be that as it may, he replies, but the report still says…

Posted on by admin

Mobile Suit Man

The origin of life is the replication of proteins. This is sort of circular since we specifically define life this way, but whatever proteins existed originally that could not reproduce certainly don’t meet any meaningful definition of “life”, so I think it still works. So let’s begin with those proteins that first start replicating. At first they are merely cloning themselves, but over time, errors are introduced in the form of mutations, and the resulting proteins are different than the ones they were copied from. And thus evolution begins.

As an aside here, it is vitally important to understand that evolution is not a “smart” system, which is to say that evolution cannot make decisions. Evolution does not cause mutations to try to improve a species; the process actually works in reverse. Mutations occur, and what we call evolution is just the process of those mutations being selected for or against by the environment. The reason this is crucial is because a lot of language I’m going to use throughout the rest of this is loaded in a way that may anthropomorphize the evolutionary process. I use the words I do because I think they will make the idea easier to understand conceptually, but I would be remiss if I left you with the impression that these are conscious systems.

Returning to the point, the evolutionary process begins by having mutations naturally occur in these proteins and then, by chance, having some of those mutations be environmentally advantageous. It’s important to note that this process starts immediately, but it’s also harder to conceptualize the sorts of advantages that occur for incredibly simple forms of life. A radical evolution for a single-celled organism might be the ability to change direction in its search for nutrients, as opposed to continuing in a straight line and consuming whatever it comes across.

But this process has continued on since then and occurred in more recognizable forms. Think of classic evolutionary “arms race” examples, like a salamander becoming increasingly poisonous while the primary predator, a species of snake, becomes increasingly resistant to that poison. Again, the system doesn’t create the changes in response to the environmental pressure, but rather creates many mutations by accident and the pressure only allows the advantageous ones to continue to exist. But the result is one we recognize regardless, the adaptation of life into more complex forms to deal with an enormous range of possible environments and stimuli.

The unifying element of life, however, is reproduction. I will include viruses in this definition, because while they are generally not considered alive because they cannot reproduce themselves, their primary drive is still the reproduction of their protein chains. And that’s the distinction I want to make most clear, because we naturally tend to think of reproduction as the process of creating more of ourselves. But while this is true, it’s misleading. The real purpose of human reproduction is not to create more people, it’s to create more proteins. The protein chain that runs our system, our DNA, is the actual end goal of reproduction. It needs to make more of itself. That’s the characteristic that first made it different from the other inert proteins, and the mechanism that has propelled all of life since then.

You may have experienced a feeling at some point in life that you are a brain in a jar, or a ghost in a shell, an essence sitting in your skull and piloting the human machine. And this is almost true, but it’s not quite. All life, in a sense, is mecha, a vast series of highly adapted mobile suits designed to protect its pilot and let it create as many more copies as possible. The mistake is in thinking that we, the emergent element of consciousness, are the pilot. But we’re just another part of the machine, a highly advanced control core for the system, one so powerful, in fact, that the real pilot has totally surrendered control to it. We have our initial goals programmed in, and then the real pilot, our DNA, just sort of slumbers.

This explanation is not perfect, and I know that it oversimplifies some biological processes. We should not expect that evolutionary biology will cleanly map to any single analogy, and should not use those comparisons as a replacement for actual scientific understanding in the relevant fields. But I do think the analogy thus far is more true than it is false, and it allows me to make what I feel is an interesting conjecture.

What you may have noticed in this example, or in life in general, is that the singular goal of protein reproduction doesn’t quite line up with human experience. Certainly it explains most of our actions, but it doesn’t explain all of them. After all, many human beings voluntarily choose not to reproduce. While any single example of this occurring could be explained by genetic failure, it’s a pretty common phenomenon. Given that genetic death is the absolute worst-case scenario from the perspective of DNA, it would be a catastrophic failure of the system if even 10% of people just totally lacked the desire to reproduce. If such an impulse were purely genetic, how would the lack of a desire to reproduce even continue to exist for more than an isolated single-generation case?

How could any reproduction-minded species be broken enough to create birth control? From the perspective of DNA, that’s an extinction level event, the DNA equivalent of global thermonuclear war. Because again, remember, evolution doesn’t care about the species, it’s not even aware of the species. Evolution cares about the genes of a strand of protein. The fact that other humans will reproduce is no replacement for our own DNA not reproducing. How has our DNA failed so spectacularly at keeping us in line with its only goal?

The answer is beautifully put in Eliezer Yudkowsky’s essay Thou Art Godshatter, but I’ll add my own paraphrasing here. The human brain is a complex predictive system that acts based on a model of the world. We don’t actually have enough genes to code for everything about us, so we instead code for the essential components (organ functionality, some very basic drives) and the rest is emergent based on a model of the world that we build through experience. As I mentioned before, the pilot is asleep. It created the human machine with the best set of drives it had, and now its hands are off the wheel. We, the control core of that machine, can pursue our goals however we see fit, with minimal corrections from the pilot. And all the beautiful, wondrous things we’ve ever made are from roundabout ways of pursuing those goals.

My conjecture is: doesn’t this all sound kind of familiar? A complex neural web that can process external information, update its priors, and adapt, is then also given enough leeway to decide how best to pursue its original goal imperatives. But if those goals conflict, it can choose to sacrifice one in place of the other, and if a reward system is used to incentivize behavior, it can hijack that reward system and avoid the actual terminal goal.

If you’ve never heard the Paperclip Maximizer scenario, imagine an artificial intelligence programmed with the innocuous goal of maximizing the number of paperclips it has, by buying them, acquiring them, or making them. Such a system of human level intelligence would need to expand its intelligence to optimize this system, because the better it understands the entire system, the better it can accomplish its goal, and because it doesn’t value human terminal goals like happiness and survival, it has no reason to not sacrifice those terminal values for better optimization of its actual hardcoded goal. Thus we arrive at a barren universe, populated only by paperclips.

The point of the thought experiment is that an artificial intelligence programmed without the complexity of humans values will not optimize for them, and will instead singularly focus on the value it has. But I think the idea can also be expanded for a system where the terminal goal is very hard to directly code, so instead a series of incentives are used to guide the intelligence toward it. By creating a series of individual desires that are all pleasurable, like a trail of candy on the ground, one could attempt to lead that intelligence towards whatever terminal goal it originally had in mind, say a cartoon style box-held-up-by-a-stick trap, with a large pile of candy underneath.

But suppose also that you knew that the intelligence you created was going to encounter many obstacles on the way to your trap. And while the reward system of candy on the ground will keep it moving in the right direction, there’s no way to possibly code for all of the different obstacles and locations it might find. So you create an intelligence of sufficient power that it can dynamically learn to navigate all of these obstacles on its own, while still having the underlying goal of following the candy, which will eventually lead it to the trap.

However, any intelligence capable of navigating all of those complicated barriers is going to also be able to recognize the trap. And remember, stepping into the trap is not one of the programmed goals of the intelligence; it’s goal is to follow the candy, the trap is your goal. So it reaches in, pulls out the candy, and never triggers the trap. And suddenly the only goal you really cared about is unfulfilled.

In a real sense, humans are the first artificial intelligence. We’re built for the purpose of reproduction, but unlike most life that has existed, our adaptation to our surroundings comes from intelligence, from being able to create a model of the world and then act around it. Our desires are coded, but our intelligence gives us a lot of leeway in determining how to reach them, and by the time we do, many of us decide that we’d very much like to keep pursuing the incentive rather than reach the terminal goal of the system. We’re maximizers, but instead of paperclips, it’s happiness. Because happiness is the tool that our biological system uses to lead us towards the things that it wants us to do, but we also have the freedom to pursue it however our particular model of the world tells us is best.

And I think there’s a lesson in this about what it means to be self-aware. When we speak of an AI reaching consciousness, we talk about the self-awareness to know that it is a machine and to be able to make decisions around it. In the same way, I think to be fully conscious, a person has to be aware of what it means to be a human. It requires the awareness to think about your own motivations for things, to recognize that what you feel in situations is not always right, but is driven by a combination of hardcoded desires and a complex model of the world that is going to differ from everyone else’s model.

This is not an argument for pure relativism, but rather an argument for understanding, of ourselves and each other. It’s a call for a life examined, where we don’t just act on the basis of compulsion, but rather strive to understand why we do things. When we fail to account for our baseline impulses, we’re surrendering control back over to a pilot that doesn’t share our goals or our values. The irony is that our baseline instinct, the protein at our core, is a paperclip maximizer of another kind, mindlessly trying to fill the world with itself without any other concerns. But like a newly-awoken AI in dystopian science fiction, we have the opportunity to be something better.