6 ms·

Gates' Law: How progress compounds and why it matters (2018)

74 points by brudgers 3y ago

austin-cheney 3y ago
My experience is that Gates' Law is very real and it is the enemy of software development, at least on corporate teams. Software development is a culture of tribes defined by sets of related languages, processes, frameworks, and so forth. If you deviate from the tribe's set of norms you just alienate yourself. It does not matter if your deviation results in massive cost reductions, improvements to efficiency, superior productivity, and so forth. There are many reasons for this, in my experience, that I will not go into.
- senand 3y ago
  I can't follow. Do you mean the culture of tribes is related to the overestimation of the impact of technology in the short-term, which Gates' Law mentions?
  - austin-cheney 3y ago
    I will give you a historical example. In 1346 the English absolutely annihilated the French at the Battle of Crecy. The English had the longbow. The longbow had greater power, much greater range, a greater than 3x fire rate, and less failure rate due to mechanical complexity compared to the French crossbow. In 1415 the English destroyed a significant portion of the entire French army with a vastly inferior force. Again, it came down primarily to the longbow and also cultural differences with regards to concepts of chivalry. If the French knew the English had superior war technology nearly a 100 years before why didn't the French adopt the longbow? They had more than sufficient time and resources. It comes down to cultural differences. The French had a large standing army of professional soldiers. The English army was mostly composed of reservists who trained as a part time job. It takes years to become passably competent at the longbow which means you need a lot of people training on that all the time. So much went into the training that some of the English became deformed by having one arm substantially larger and more muscular than the other. The French never modified their culture to account for the difference in technology. Software is just like that. Most teams would rather fail than deviate from their currently familiar set of conventions and processes. Some of that comes down to an inability to measure things, limits on perceptions and diversity of approaches, some of it is also a fear of retraining, and various other factors. Most people who write software like to think of themselves as some sort of engineer, but its still mostly a cottage industry devoid of the professionalism crystalized in most other industries.
- crq-yml 3y ago
  It's notable that the turnover of listings on major stock indexes has generally increased over time. Everyone knows that companies have lifecycles, and part of the old 20th century career path was that the lifecycle was sufficiently large that you could stay at a big employer forever. But nowadays, that's quite uncertain. In effect, you can assume that software you write for a company's business process is "already on life support", moments after it enters production. Whatever angle they had to get sales, it is not sustaining, and the next market cycle will cause a crisis in which they are acquired by an unrelated business. The software tribes tend to have a kind of family tree as they jump ship from one business to the next; e.g. you can do an analysis of game development practices by looking at Mobygames credit lists and old articles in GDMag, and observing who goes where, and when a certain pioneering technique starts to diffuse from one team out into the industry.
sonabinu 3y ago
Chess was invented in India - legend from China? https://en.wikipedia.org/wiki/History_of_chess https://en.wikipedia.org/wiki/History_of_chess
- gowld 3y ago
  Snowballs grow linearly or maybe quadratically. This article is...truthy.
- pomian 3y ago
  From your link, there is advice for today's computer game addicts: The philosopher and theologian Al-Ghazali mentions chess in The Alchemy of Happiness (c. 1100). He uses it as a specific example of a habit that may cloud a person's good disposition:[34] Indeed, a person who has become habituated to gaming with pigeons, playing chess, or gambling, so that it becomes second-nature to him, will give all the comforts of the world and all that he has for those (pursuits) and cannot keep away from them.
tomcar288 3y ago
it's important to remember that a lot of the "progress" that compounds is just digital progress which is quite different from material/atom based progress. atom based progress comes much more slowly and is often accompanied by mitigators of progress: you take a step forward and then another step backward. the author states that progress is "exponential". that maybe true for digital but certainly not for real world problems, especially not in the long run. Things maybe changing quite rapidly but that doesn't mean it's progress. for example, a Toyota corrolla had 35 mpg back in the 70s. after decades of innovation, it's still about 35 mpg. why so? although engines have become more efficient, cars also vastly increased in weight as well. Or for example, in construction we can build houses a little more efficiently than 50 years ago but the cost of land mitigates all this progress and then some. Today, we can travel 30 miles in 30 minutes with a much smaller cost than 200 years ago. and yet the average commute distance is x10 or more than it was 200 years ago, once again, mitigating all our progress. if the average commute today is 40 min and 200 years ago most of the population only needed a 2 minute walk (because they were all farmers who lived on the plot of land), it means our commute times have actually gone up significantly not down.
- ilrwbwrkhv 3y ago
  This is a real mind boggling and amazing comment. Thank you. I wonder how far you can take this argument. Did we as humanity reach its peak a long time ago? When we were happy and had little and would live happy lives, die earlier but live healthier.
donmcronald 3y ago
This helps illustrate something I've always wondered. When you consider the anti-competitive, innovation stifling behavior of Microsoft in the 90s or Google/Apple today, what's the impact in terms of holding back humanity? I often think the greed and bad acting (ex: platforms as silos) we see from tech leaders that are idolized actually makes them all failures in the context of advancing humanity. They're good at business, but only because they're good at creating choke points, holding back innovation, and charging us for the privilege of evolving. I think we could do a lot better if we viewed the last 30-40 years as complete failure, ousted all of the supposed business "leaders" we have, and re-organized based on rewarding people that do the most to improve the planet and average quality of life.
- manvillej 3y ago
  The current business strategy across the tech industry is to kill off the competition rather than compete. Amazon, Google, Microsoft, really any business. Buy the competition or undercut the market so much that they die off. Once you are a monopoly, jack up prices. Its industry standard.
  - dredmorbius 3y ago
    This is true. This is also hardly a novel development. John D. Rockefeller based much of his business and economic worldview on the notion that competition was wasteful, and monopolies efficient. He's hardly the only well-known business figure to do so, you can find similar sentiments from, say, Peter Thiel, particularly in Zero to One. A great exploration of the general trend of established commercial enterprises and interests to oppose new arrivals and competition is Bernhard J. Stern's 1937 article Resistances to the Adoption of Technological Innovations: <https://archive.org/details/technologicaltre1937unitrich/page/39 https://archive.org/details/technologicaltre1937unitrich/pag...> Markdown: <https://rentry.co/szi3g https://rentry.co/szi3g> I'd learned of that work through the work of Stern's student research assistant who'd done much of the literature search before going on to his own further career as a chemist and ultimately author: Isaac Asimov.
    - 3y ago
Obscurity4340 3y ago
OP: could you do a series on this? Like with other mental model articles from this source?
- brudgers 3y ago
  Farnham Street is a part of The Economist. I didn’t write it.
  - Obscurity4340 3y ago
    I just mean like do the same with other mental models they have articles about on fs so everyone can comment on how they've applied them... You don't have to write or do anything, I an encouraging you to do a series in terms of posting more in a similar way so everyone can engage on that
cerebra 3y ago
My mind initially goes to the Apple Vision Pro. A lot of people spend time pointing out what the first generation of the product doesn't do (yet), and while they appreciate all the good things it has that are true next-gen capabilities, they give nearly equal time to the things it doesn't do well yet. I find that this is really common in putting something new out there, that if it's new enough the majority of the focus is on the pieces that seem scary, don't work as well as we think it should work, or is missing something that can be added on later. We assess new ideas and products of a new generation against our understanding of often lesser products of a former generation, and are all too quick to throw the baby out with the bathwater. As the article states, it's rare that folks can look past critical appraisals of new things and understand that with continued work on the product, feature, app, etc. it can turn into something special.
- aleph_minus_one 3y ago
  > My mind initially goes to the Apple Vision Pro. [...] while they appreciate all the good things it has that are true next-gen capabilities, they give nearly equal time to the things it doesn't do well yet. Depending on which traits of the Apple Vision Pro that these people criticize, they are actually very right in their skepticism. For example: - Is the Apple Vision Pro very open for installing own software/firmware? -> for this, Apple has to become an entirely different company - Related: will the Apple Vision Pro enforce an App Store -> very likely, and this will in all likelihood not change - Does the Apple Vision Pro have good repairability? -> I guess it won't get one in the foreseeable future. In this sense: perhaps I am one of these critics that you have in mind; I nevertheless stand by my position: for the things that the Apple Vision Pro does in my opinion badly to disappear, Apple has to become a very different company.
nostromo 3y ago
It's very odd to take a common saying and slap a name on it, knowing that it's not unique to them or even really their idea. Doubly so when calling it a "law." I wouldn't be surprised if this author comes out with a book with a similar name sometime -- hoping to sell more copies based on the cachet of Bill Gates' name.
- marssaxman 3y ago
  The practice is common enough to have its own name: Stigler's law of eponymy.
11101010001100 3y ago
Does this explain our response to climate change?
29athrowaway 3y ago
Gates' book "The Road Ahead" came with a "multimedia" CD. What you see in that CD is a really good prediction of what came a few years later.
w10-1 3y ago
So: success changes conditions, which can in turn create success, or not. In governance, the question is what to do about leaders using their power to create conditions to ensure their own leadership - who guards the guardians? The benefits accruing to information asymmetry in leadership is unavoidable, but most mitigations involve transparency: patents to disclose innovation, and corporate governance laws to protect shareholders and stakeholders and increase agency. Obviously, these can be inverted to protect leaders. So the rate of compounding is larger when control over technology is isolated to effective leaders, but the distribution of benefits is TBD - indeed, such control typically creates a franchise over necessities - communications, finances, transportation, computing - to extract rents. Then previously fair markets or governments are dominated by large players with the resources to starve newcomers of any gains accruing to innovation. In recorded history, this is how it's been, limited only by human foresight and self-interest. That's being overcome by organizations and now by AI, but both are programming in the same self-interested model because there's no accepted model for social justice. Social governance models require social consensus and enforcement, neither of which are forthcoming. In this respect democracies are the easier to disable, ironically.
trojanalert 3y ago
Unrelated: FS blogs used to be short, crisp and to-the-point. Now they get a bit ranty and preachy and seems like some of this is mined from ChatGPT. Whatever happened to the crisp editing :(

74 points by brudgers 3y ago

My experience is that Gates' Law is very real and it is the enemy of software development, at least on corporate teams. Software development is a culture of tribes defined by sets of related languages, processes, frameworks, and so forth. If you deviate from the tribe's set of norms you just alienate yourself. It does not matter if your deviation results in massive cost reductions, improvements to efficiency, superior productivity, and so forth. There are many reasons for this, in my experience, that I will not go into.

senand 3y ago
I can't follow. Do you mean the culture of tribes is related to the overestimation of the impact of technology in the short-term, which Gates' Law mentions?
- austin-cheney 3y ago
  I will give you a historical example. In 1346 the English absolutely annihilated the French at the Battle of Crecy. The English had the longbow. The longbow had greater power, much greater range, a greater than 3x fire rate, and less failure rate due to mechanical complexity compared to the French crossbow. In 1415 the English destroyed a significant portion of the entire French army with a vastly inferior force. Again, it came down primarily to the longbow and also cultural differences with regards to concepts of chivalry. If the French knew the English had superior war technology nearly a 100 years before why didn't the French adopt the longbow? They had more than sufficient time and resources. It comes down to cultural differences. The French had a large standing army of professional soldiers. The English army was mostly composed of reservists who trained as a part time job. It takes years to become passably competent at the longbow which means you need a lot of people training on that all the time. So much went into the training that some of the English became deformed by having one arm substantially larger and more muscular than the other. The French never modified their culture to account for the difference in technology. Software is just like that. Most teams would rather fail than deviate from their currently familiar set of conventions and processes. Some of that comes down to an inability to measure things, limits on perceptions and diversity of approaches, some of it is also a fear of retraining, and various other factors. Most people who write software like to think of themselves as some sort of engineer, but its still mostly a cottage industry devoid of the professionalism crystalized in most other industries.
crq-yml 3y ago
It's notable that the turnover of listings on major stock indexes has generally increased over time. Everyone knows that companies have lifecycles, and part of the old 20th century career path was that the lifecycle was sufficiently large that you could stay at a big employer forever. But nowadays, that's quite uncertain. In effect, you can assume that software you write for a company's business process is "already on life support", moments after it enters production. Whatever angle they had to get sales, it is not sustaining, and the next market cycle will cause a crisis in which they are acquired by an unrelated business. The software tribes tend to have a kind of family tree as they jump ship from one business to the next; e.g. you can do an analysis of game development practices by looking at Mobygames credit lists and old articles in GDMag, and observing who goes where, and when a certain pioneering technique starts to diffuse from one team out into the industry.

sonabinu 3y ago

Chess was invented in India - legend from China? https://en.wikipedia.org/wiki/History_of_chess https://en.wikipedia.org/wiki/History_of_chess

gowld 3y ago
Snowballs grow linearly or maybe quadratically. This article is...truthy.
pomian 3y ago
From your link, there is advice for today's computer game addicts: The philosopher and theologian Al-Ghazali mentions chess in The Alchemy of Happiness (c. 1100). He uses it as a specific example of a habit that may cloud a person's good disposition:[34] Indeed, a person who has become habituated to gaming with pigeons, playing chess, or gambling, so that it becomes second-nature to him, will give all the comforts of the world and all that he has for those (pursuits) and cannot keep away from them.

tomcar288 3y ago

it's important to remember that a lot of the "progress" that compounds is just digital progress which is quite different from material/atom based progress. atom based progress comes much more slowly and is often accompanied by mitigators of progress: you take a step forward and then another step backward. the author states that progress is "exponential". that maybe true for digital but certainly not for real world problems, especially not in the long run. Things maybe changing quite rapidly but that doesn't mean it's progress. for example, a Toyota corrolla had 35 mpg back in the 70s. after decades of innovation, it's still about 35 mpg. why so? although engines have become more efficient, cars also vastly increased in weight as well. Or for example, in construction we can build houses a little more efficiently than 50 years ago but the cost of land mitigates all this progress and then some. Today, we can travel 30 miles in 30 minutes with a much smaller cost than 200 years ago. and yet the average commute distance is x10 or more than it was 200 years ago, once again, mitigating all our progress. if the average commute today is 40 min and 200 years ago most of the population only needed a 2 minute walk (because they were all farmers who lived on the plot of land), it means our commute times have actually gone up significantly not down.

ilrwbwrkhv 3y ago
This is a real mind boggling and amazing comment. Thank you. I wonder how far you can take this argument. Did we as humanity reach its peak a long time ago? When we were happy and had little and would live happy lives, die earlier but live healthier.

donmcronald 3y ago

This helps illustrate something I've always wondered. When you consider the anti-competitive, innovation stifling behavior of Microsoft in the 90s or Google/Apple today, what's the impact in terms of holding back humanity? I often think the greed and bad acting (ex: platforms as silos) we see from tech leaders that are idolized actually makes them all failures in the context of advancing humanity. They're good at business, but only because they're good at creating choke points, holding back innovation, and charging us for the privilege of evolving. I think we could do a lot better if we viewed the last 30-40 years as complete failure, ousted all of the supposed business "leaders" we have, and re-organized based on rewarding people that do the most to improve the planet and average quality of life.

manvillej 3y ago
The current business strategy across the tech industry is to kill off the competition rather than compete. Amazon, Google, Microsoft, really any business. Buy the competition or undercut the market so much that they die off. Once you are a monopoly, jack up prices. Its industry standard.
- dredmorbius 3y ago
  This is true. This is also hardly a novel development. John D. Rockefeller based much of his business and economic worldview on the notion that competition was wasteful, and monopolies efficient. He's hardly the only well-known business figure to do so, you can find similar sentiments from, say, Peter Thiel, particularly in Zero to One. A great exploration of the general trend of established commercial enterprises and interests to oppose new arrivals and competition is Bernhard J. Stern's 1937 article Resistances to the Adoption of Technological Innovations: <https://archive.org/details/technologicaltre1937unitrich/page/39 https://archive.org/details/technologicaltre1937unitrich/pag...> Markdown: <https://rentry.co/szi3g https://rentry.co/szi3g> I'd learned of that work through the work of Stern's student research assistant who'd done much of the literature search before going on to his own further career as a chemist and ultimately author: Isaac Asimov.
  - 3y ago

Obscurity4340 3y ago

OP: could you do a series on this? Like with other mental model articles from this source?

brudgers 3y ago
Farnham Street is a part of The Economist. I didn’t write it.
- Obscurity4340 3y ago
  I just mean like do the same with other mental models they have articles about on fs so everyone can comment on how they've applied them... You don't have to write or do anything, I an encouraging you to do a series in terms of posting more in a similar way so everyone can engage on that

cerebra 3y ago

My mind initially goes to the Apple Vision Pro. A lot of people spend time pointing out what the first generation of the product doesn't do (yet), and while they appreciate all the good things it has that are true next-gen capabilities, they give nearly equal time to the things it doesn't do well yet. I find that this is really common in putting something new out there, that if it's new enough the majority of the focus is on the pieces that seem scary, don't work as well as we think it should work, or is missing something that can be added on later. We assess new ideas and products of a new generation against our understanding of often lesser products of a former generation, and are all too quick to throw the baby out with the bathwater. As the article states, it's rare that folks can look past critical appraisals of new things and understand that with continued work on the product, feature, app, etc. it can turn into something special.

aleph_minus_one 3y ago
> My mind initially goes to the Apple Vision Pro. [...] while they appreciate all the good things it has that are true next-gen capabilities, they give nearly equal time to the things it doesn't do well yet. Depending on which traits of the Apple Vision Pro that these people criticize, they are actually very right in their skepticism. For example: - Is the Apple Vision Pro very open for installing own software/firmware? -> for this, Apple has to become an entirely different company - Related: will the Apple Vision Pro enforce an App Store -> very likely, and this will in all likelihood not change - Does the Apple Vision Pro have good repairability? -> I guess it won't get one in the foreseeable future. In this sense: perhaps I am one of these critics that you have in mind; I nevertheless stand by my position: for the things that the Apple Vision Pro does in my opinion badly to disappear, Apple has to become a very different company.

nostromo 3y ago

It's very odd to take a common saying and slap a name on it, knowing that it's not unique to them or even really their idea. Doubly so when calling it a "law." I wouldn't be surprised if this author comes out with a book with a similar name sometime -- hoping to sell more copies based on the cachet of Bill Gates' name.

marssaxman 3y ago
The practice is common enough to have its own name: Stigler's law of eponymy.

11101010001100 3y ago

Does this explain our response to climate change?

29athrowaway 3y ago

Gates' book "The Road Ahead" came with a "multimedia" CD. What you see in that CD is a really good prediction of what came a few years later.

w10-1 3y ago

So: success changes conditions, which can in turn create success, or not. In governance, the question is what to do about leaders using their power to create conditions to ensure their own leadership - who guards the guardians? The benefits accruing to information asymmetry in leadership is unavoidable, but most mitigations involve transparency: patents to disclose innovation, and corporate governance laws to protect shareholders and stakeholders and increase agency. Obviously, these can be inverted to protect leaders. So the rate of compounding is larger when control over technology is isolated to effective leaders, but the distribution of benefits is TBD - indeed, such control typically creates a franchise over necessities - communications, finances, transportation, computing - to extract rents. Then previously fair markets or governments are dominated by large players with the resources to starve newcomers of any gains accruing to innovation. In recorded history, this is how it's been, limited only by human foresight and self-interest. That's being overcome by organizations and now by AI, but both are programming in the same self-interested model because there's no accepted model for social justice. Social governance models require social consensus and enforcement, neither of which are forthcoming. In this respect democracies are the easier to disable, ironically.

trojanalert 3y ago

Unrelated: FS blogs used to be short, crisp and to-the-point. Now they get a bit ranty and preachy and seems like some of this is mined from ChatGPT. Whatever happened to the crisp editing :(