Great write-up. Several things came to mind as I read this:
- Peter Thiel would nod in approval.
- The ghost of Jack Welch is shaking his ethereal fist at you.
- It's weird how different companies feel compelled to slap their own label on something. I think (can't confirm) that Jack Welch called the 20-70-10 system a "Vitality curve" but Microsoft insisted on calling essentially the same thing "stack ranking."
- Have you heard of the blue ocean strategy? It was popular a few years ago, but has fallen out of fashion lately. The idea is to hoist yourself out of oceans that run red with the blood of competition, immerse yourself in blue oceans untainted by competitive pressures.
- Steve Ballmer was new at the helm when he implemented stack ranking at Microsoft, and Bill Gates had stepped down because of a massive anti-competition lawsuit from various regulators. That background provides some context for (but doesn't necessarily justify) the choice to implement stack ranking.
> competition as the only way to drive the world forwards
In a certain, very important, sense, it is indeed. Evolution is a competitive system, in fact it's the system that M$'s sociopathic reward system is a crude imitation of. It's demonstrably, objectively true that "Stack Ranking" (as implemented by Nature) does indeed hill-climb to the highest local maxima attainable by a species.
What people usually fail to consider is :
1- Moral Cost : After all it would be extremly efficient if we could simply mass-murder all people with heritable diseases, we would probably be saving a whole (exponentially-growing) tree of descendents from diseases too if we did it before they have children.
Evolution is an undirected undesigned amoral process, so it doesn't feel shame, remorse, guilt, etc... But we humans are goal-directed, morality-constrained, design-seeking moral agents, so we do feel shame and guilt and remorse, and there is nothing more shameful than treating other humans as a tool.
Human-Designed systems has to grapple with the is-ought distinction, and has to defend the (often implicit) claims that its design makes. Implicit in M$'s reward system is the claim "making money for the company is so extremly important, so vital and non-negotiable, that this is worth the Moral Cost associated with it".
This is wrong : Making money for M$ is actually so utterly without value, so unimportant and banal, that it's not worth any moral cost at all, let alone the one associated with such a cruel system.
But consider an alternate 3-body-problem-like world where employees in a company are actually working around the clock to make more and better spaceships for humanity to escape/fight aliens, aliens who are approaching Earth and has the explicit intent to murder everybody on it. In this world, is it actually worth it to have "Stack Rankings" ? I would say yes. Perhaps with some modifications, like (a) Making the criteria for evaluations extremly objective and personal-judgement-independent (b) Making the evaluators a random and constantly-changing subset of the top 1% of those who survived the last N evaluations (c) Make it so that everybody has a non-zero, greater-than-10% chance of being fired, including the very top of the organization.
In short, Competitive Systems, even very cruel ones like Evolution, work. Your own existence is a living walking proof of this. They just have a very high negative moral cost associated with them, a cost that can still be paid if the goal we're working toward is sufficiently worthy.
2- The Outcome Needn't Be Death Or Starvation : Unlike Evolution, we have a whole selection of outcomes that we can assert for those who fall behind in Competition. We can have UBI-ish safety nets so that work is never the precondition for eating food and sleeping in beds, we can have a culture that doesn't attach self-worth to work (or, more accurately, **Jobs**), and considered them on an equal footing with hobbies, unpaid volunteer work, unpaid house work, unpaid open source programming, etc...
3- The Process Needn't Be Cruel : This is a continuation of point (2), and point (2) is a precondition of it (i.e. a process whose outcome is death or starvation is cruel by definition). People can always be compensated extra for being fired, promised assistance till they find the next job, reviewers can be anonymous and required to be respectful (where the defintion of respectful is in the control of fired people, and if enough people complain the reviewer is barred from ever reviewing again.) In short, signal that people are valuable. Signal that <doing bad thing> to them was hard, the least-evil alternative from a set of more evil choices.
tldr; your view about competition is not nuanced enough, unless you are just baiting/trolling engagement (in which case congrats, I guess.). There is someone who spends a lot of time thinking about Competition and when it's a Good Thing (^TM), that person is Liv Boeree, start with her Moloch: The Media Wars video (https://www.youtube.com/watch?v=PRz54V7rU4U), and maybe see a bunch of her interviews and podcasts. The key message is that Competition is not uniformly bad, it has costs and externalities, those can be evaluated and summed, and if the final sum is positive then the competition system is worth it.
First of all thank you for your thought out reply!
I think I agree with you that once the stakes of an endeavour are high enough, it's ok to prioritize top performance over other considerations (also I think we both agree nothing MS was working on fits this category to justify such behaviour) - but even in cases such as this, something like the Stack System literally wouldn't work.
Not due to moral reasons, but for the strictly practical reasons mentioned in the post - the myriad of negative outcomes that arise from establishing such perverse incentives on a team are catastrophic to progress (even considering the alterations you suggest above).
Also it's important to note the emphasis at the end of the post: It's unquestionable that both approachs "work", the question is what approach has the better cost:reward tradeoff (as you rightly pointed out yourself!). Positive sum frameworks will by definition always beat zero sum or negative sum ones. I'm not baiting anything! This is sincere.
To take your space race example - right now it's SpaceX and Boeing competing against eachother to reach Mars first. In what scenario do you think humanity reaches Mars faster: If competition drives the pace of each company's work and innovation in an attempt to one up eachother, or if both pooled their joint resources and tried to solve the problem together? The same principle applies to interpersonal collaboration too.
Also I'll watch the link you sent! Thank you for that!
Thanks very much for your kind words, it's always rewarding when people like my ramblings on things haha.
> if both pooled their joint resources and tried to solve the problem together?
Don't you agree that Cooperation (i.e. "pooling resources") has limits ? This is the entire reason why any multiplayer human institution fails, and the original sin of all Civilization. Humans are simply built to cooperate within 100-1000 person sized groups of blood-related-ish people, the extreme and gargantuan scale that modern human groups operate within is unimaginable to our ancestors (whom we share the same brains with), and this scale stretches our social protocols to the absolute limit, in the same way that cars travelling at 150 Km/h are unimaginable to them and they stretch our ancient eyes and reflexes to the absolute limit.
There are a lot of hacks you can do to cope with that fundamental limit, sure. One way is Hierarchy, which is just Divide-And-Conquer problem-solving applied to human organization. You can also invent Gods, Laws, Monarchy, the concept of a "Stock Market", Money, etc... All of those are imperfect tools for Cooperation, none of them solve the problem satisfactorily, and when they break the consequences are often catastrophic.
Competition is just another tool for solving the problem of mass-scale human coordination. It initially requires a minimal amount of Cooperation between the competing entities to agree on the rules of competition and other groundwork, but otherwise it just acknowledges that Cooperation is hard and so simply doesn't attempt it at all.
In other words, Competition is a healthy alternative when Cooperation/Management fails (which it must, because humans weren't designed for civilization and the massively-sized groups it involves). SpaceX and Boeing are barely able to manage themselves as it is, just like any extremely large corporations, they are no doubt full of bullshit and time-eating pointless meetings and management "stand-ups" just to get everyone on the same page, imagine if the 2 giants merged and some poor bastard(s) had to manage it all. It's simply much more efficient to let both "duplicate" work than to pay the huge penalty of synchronization.
In addition - and independently of the above point - Competition implies Robustness and Parallel Problem Solving. Meaning : even if Boeing and SpaceX *could* be managed under a single leadership, that would imply that all the mistakes of that leadership would impact the value of both company's resources. Competition means many eggs in multiple baskets. Even if there are no mistakes and no stupid leaderships, Competition engenders multiple and parallel solutions to the same problems, which is desirable almost everywhere, as all hard problems have different solutions with different tradeoffs.
> Positive sum frameworks will by definition always beat zero sum or negative sum ones
This only holds if you're ignoring Communication costs, eh ? If more Cooperation was always good then multicellular life would have out-evolutioned unicellular life long ago (and viruses, which I'm told that biologists aren't even sure are proper living organisms), but that's not what happened. If more Cooperation was always good then multicellular life would keep growing and growing till the whole biomass of the planet is a single organism, but that's not what happened.
Why ? Because Cooperation has costs, and that communication cost scales quadratically in the number of cooperating entities. Cooperation brings with it so many opportunities for misunderstandings, duplicated work, free-riders, and indeed even Competition, except it's a backstabbing kind of bad competition, the kind that formal explicit competitions try to avoid. The more numerous and "distant" (meaning no shared culture, blood, destiny, etc... anything to bind them together) are the entities who must cooperate, the harder and less efficient their cooperation is. There comes a point - different for each group - where even crude Free-For-All Competition is more efficient than Cooperation.
There are a lot of tradeoffs and fuzzy limits involved, I don't claim to know when and where is that exact point when Competitions beats Cooperation, I just claim it exists, it exists for any group of cooperating agents/entities, and most giant organizations or institutions are actually beyond it and would benefit from being broken up into competing parts.
> tldr; your view about competition is not nuanced enough
I didn't get the impression that OP was saying "competition and natural selection = bad." I interpreted it as a case study of "here's one case where ruthless competition caused all kinds of unintended consequences in the long run."
You could, hypothetically, lay out the a fully-fleshed out case in favor of monopoly vs. competition, but that's basically an entire treatise by itself. Peter Thiel already did that in one of the chapters of Zero to One. Substack isn't well-suited to publishing treatises, though a series of post could be compiled together into a coherent treatise on a topic, showing all facets of a complex problem.
Definitely makes more sense. I'm always prefer the approach of dedicating resources to helping employees develop their competence rather than firing X% once a year - but of-course it's unavoidable that sometimes people need to be let go.
What the fuck was that first sentence? I stopped reading immediately.
Thank you for your feedback Fart 😂
I suspected it was a high risk one and here we are.
Also thank you for subscribing! I appreciate you bearing the risk that my sentence structures will annoy you again in the future.
I found it rather evocative! Drew me in at least
I could imagine Peter Thiel bobbing his head in approval as he read that first sentence.
Haha cheers Sai! I appreciate you taking the time to say that!
Did you even read the sentence?
Of course they did! Did you? At the very least there is an extraneous comma.
reletavistic --> relativistic
Briliant.
:')
Great write-up. Several things came to mind as I read this:
- Peter Thiel would nod in approval.
- The ghost of Jack Welch is shaking his ethereal fist at you.
- It's weird how different companies feel compelled to slap their own label on something. I think (can't confirm) that Jack Welch called the 20-70-10 system a "Vitality curve" but Microsoft insisted on calling essentially the same thing "stack ranking."
- Have you heard of the blue ocean strategy? It was popular a few years ago, but has fallen out of fashion lately. The idea is to hoist yourself out of oceans that run red with the blood of competition, immerse yourself in blue oceans untainted by competitive pressures.
- Steve Ballmer was new at the helm when he implemented stack ranking at Microsoft, and Bill Gates had stepped down because of a massive anti-competition lawsuit from various regulators. That background provides some context for (but doesn't necessarily justify) the choice to implement stack ranking.
> competition as the only way to drive the world forwards
In a certain, very important, sense, it is indeed. Evolution is a competitive system, in fact it's the system that M$'s sociopathic reward system is a crude imitation of. It's demonstrably, objectively true that "Stack Ranking" (as implemented by Nature) does indeed hill-climb to the highest local maxima attainable by a species.
What people usually fail to consider is :
1- Moral Cost : After all it would be extremly efficient if we could simply mass-murder all people with heritable diseases, we would probably be saving a whole (exponentially-growing) tree of descendents from diseases too if we did it before they have children.
Evolution is an undirected undesigned amoral process, so it doesn't feel shame, remorse, guilt, etc... But we humans are goal-directed, morality-constrained, design-seeking moral agents, so we do feel shame and guilt and remorse, and there is nothing more shameful than treating other humans as a tool.
Human-Designed systems has to grapple with the is-ought distinction, and has to defend the (often implicit) claims that its design makes. Implicit in M$'s reward system is the claim "making money for the company is so extremly important, so vital and non-negotiable, that this is worth the Moral Cost associated with it".
This is wrong : Making money for M$ is actually so utterly without value, so unimportant and banal, that it's not worth any moral cost at all, let alone the one associated with such a cruel system.
But consider an alternate 3-body-problem-like world where employees in a company are actually working around the clock to make more and better spaceships for humanity to escape/fight aliens, aliens who are approaching Earth and has the explicit intent to murder everybody on it. In this world, is it actually worth it to have "Stack Rankings" ? I would say yes. Perhaps with some modifications, like (a) Making the criteria for evaluations extremly objective and personal-judgement-independent (b) Making the evaluators a random and constantly-changing subset of the top 1% of those who survived the last N evaluations (c) Make it so that everybody has a non-zero, greater-than-10% chance of being fired, including the very top of the organization.
In short, Competitive Systems, even very cruel ones like Evolution, work. Your own existence is a living walking proof of this. They just have a very high negative moral cost associated with them, a cost that can still be paid if the goal we're working toward is sufficiently worthy.
2- The Outcome Needn't Be Death Or Starvation : Unlike Evolution, we have a whole selection of outcomes that we can assert for those who fall behind in Competition. We can have UBI-ish safety nets so that work is never the precondition for eating food and sleeping in beds, we can have a culture that doesn't attach self-worth to work (or, more accurately, **Jobs**), and considered them on an equal footing with hobbies, unpaid volunteer work, unpaid house work, unpaid open source programming, etc...
3- The Process Needn't Be Cruel : This is a continuation of point (2), and point (2) is a precondition of it (i.e. a process whose outcome is death or starvation is cruel by definition). People can always be compensated extra for being fired, promised assistance till they find the next job, reviewers can be anonymous and required to be respectful (where the defintion of respectful is in the control of fired people, and if enough people complain the reviewer is barred from ever reviewing again.) In short, signal that people are valuable. Signal that <doing bad thing> to them was hard, the least-evil alternative from a set of more evil choices.
tldr; your view about competition is not nuanced enough, unless you are just baiting/trolling engagement (in which case congrats, I guess.). There is someone who spends a lot of time thinking about Competition and when it's a Good Thing (^TM), that person is Liv Boeree, start with her Moloch: The Media Wars video (https://www.youtube.com/watch?v=PRz54V7rU4U), and maybe see a bunch of her interviews and podcasts. The key message is that Competition is not uniformly bad, it has costs and externalities, those can be evaluated and summed, and if the final sum is positive then the competition system is worth it.
First of all thank you for your thought out reply!
I think I agree with you that once the stakes of an endeavour are high enough, it's ok to prioritize top performance over other considerations (also I think we both agree nothing MS was working on fits this category to justify such behaviour) - but even in cases such as this, something like the Stack System literally wouldn't work.
Not due to moral reasons, but for the strictly practical reasons mentioned in the post - the myriad of negative outcomes that arise from establishing such perverse incentives on a team are catastrophic to progress (even considering the alterations you suggest above).
Also it's important to note the emphasis at the end of the post: It's unquestionable that both approachs "work", the question is what approach has the better cost:reward tradeoff (as you rightly pointed out yourself!). Positive sum frameworks will by definition always beat zero sum or negative sum ones. I'm not baiting anything! This is sincere.
To take your space race example - right now it's SpaceX and Boeing competing against eachother to reach Mars first. In what scenario do you think humanity reaches Mars faster: If competition drives the pace of each company's work and innovation in an attempt to one up eachother, or if both pooled their joint resources and tried to solve the problem together? The same principle applies to interpersonal collaboration too.
Also I'll watch the link you sent! Thank you for that!
Thanks very much for your kind words, it's always rewarding when people like my ramblings on things haha.
> if both pooled their joint resources and tried to solve the problem together?
Don't you agree that Cooperation (i.e. "pooling resources") has limits ? This is the entire reason why any multiplayer human institution fails, and the original sin of all Civilization. Humans are simply built to cooperate within 100-1000 person sized groups of blood-related-ish people, the extreme and gargantuan scale that modern human groups operate within is unimaginable to our ancestors (whom we share the same brains with), and this scale stretches our social protocols to the absolute limit, in the same way that cars travelling at 150 Km/h are unimaginable to them and they stretch our ancient eyes and reflexes to the absolute limit.
There are a lot of hacks you can do to cope with that fundamental limit, sure. One way is Hierarchy, which is just Divide-And-Conquer problem-solving applied to human organization. You can also invent Gods, Laws, Monarchy, the concept of a "Stock Market", Money, etc... All of those are imperfect tools for Cooperation, none of them solve the problem satisfactorily, and when they break the consequences are often catastrophic.
Competition is just another tool for solving the problem of mass-scale human coordination. It initially requires a minimal amount of Cooperation between the competing entities to agree on the rules of competition and other groundwork, but otherwise it just acknowledges that Cooperation is hard and so simply doesn't attempt it at all.
In other words, Competition is a healthy alternative when Cooperation/Management fails (which it must, because humans weren't designed for civilization and the massively-sized groups it involves). SpaceX and Boeing are barely able to manage themselves as it is, just like any extremely large corporations, they are no doubt full of bullshit and time-eating pointless meetings and management "stand-ups" just to get everyone on the same page, imagine if the 2 giants merged and some poor bastard(s) had to manage it all. It's simply much more efficient to let both "duplicate" work than to pay the huge penalty of synchronization.
In addition - and independently of the above point - Competition implies Robustness and Parallel Problem Solving. Meaning : even if Boeing and SpaceX *could* be managed under a single leadership, that would imply that all the mistakes of that leadership would impact the value of both company's resources. Competition means many eggs in multiple baskets. Even if there are no mistakes and no stupid leaderships, Competition engenders multiple and parallel solutions to the same problems, which is desirable almost everywhere, as all hard problems have different solutions with different tradeoffs.
> Positive sum frameworks will by definition always beat zero sum or negative sum ones
This only holds if you're ignoring Communication costs, eh ? If more Cooperation was always good then multicellular life would have out-evolutioned unicellular life long ago (and viruses, which I'm told that biologists aren't even sure are proper living organisms), but that's not what happened. If more Cooperation was always good then multicellular life would keep growing and growing till the whole biomass of the planet is a single organism, but that's not what happened.
Why ? Because Cooperation has costs, and that communication cost scales quadratically in the number of cooperating entities. Cooperation brings with it so many opportunities for misunderstandings, duplicated work, free-riders, and indeed even Competition, except it's a backstabbing kind of bad competition, the kind that formal explicit competitions try to avoid. The more numerous and "distant" (meaning no shared culture, blood, destiny, etc... anything to bind them together) are the entities who must cooperate, the harder and less efficient their cooperation is. There comes a point - different for each group - where even crude Free-For-All Competition is more efficient than Cooperation.
There are a lot of tradeoffs and fuzzy limits involved, I don't claim to know when and where is that exact point when Competitions beats Cooperation, I just claim it exists, it exists for any group of cooperating agents/entities, and most giant organizations or institutions are actually beyond it and would benefit from being broken up into competing parts.
> tldr; your view about competition is not nuanced enough
I didn't get the impression that OP was saying "competition and natural selection = bad." I interpreted it as a case study of "here's one case where ruthless competition caused all kinds of unintended consequences in the long run."
You could, hypothetically, lay out the a fully-fleshed out case in favor of monopoly vs. competition, but that's basically an entire treatise by itself. Peter Thiel already did that in one of the chapters of Zero to One. Substack isn't well-suited to publishing treatises, though a series of post could be compiled together into a coherent treatise on a topic, showing all facets of a complex problem.
Jeeze---ten percent every three months?
I think I heard the guys on All In talk about a company that fired the bottom 5% each year and say that sounded good. I didn't agree until now.
Definitely makes more sense. I'm always prefer the approach of dedicating resources to helping employees develop their competence rather than firing X% once a year - but of-course it's unavoidable that sometimes people need to be let go.
Someone should tell this story to SalesForce.
Incredible how truly, incompetently stupid the idea was. Of course a dumbfuck fail-upwards blowhard like Ballmer was all for it.
I still remember the mental whiplash I got the first time I read the phrase "Failing Upwards". Suddenly so much in the world clicked and made sense!