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

by brudgers • 2 years ago

74 44 2 years 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• 2 years 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-cheney2• 2 years 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• 2 years 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• 2 years ago

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

gowld• 2 years ago

Snowballs grow linearly or maybe quadratically.

This article is...truthy.

pomian• 2 years 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• 2 years 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• 2 years 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.

donmcronald4• 2 years 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• 2 years 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.

dredmorbius2• 2 years 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/pag...>

Markdown: <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.

anthonypasq• 2 years ago

re-organized based on rewarding people that do the most to improve the planet and average quality of life

Any centralized human attempt to do this has generally been a disaster, and a capitalist would argue that markets are you best shot at doing this.

donmcronald4• 2 years ago

I would tend to agree. Oust was probably too strong of a word. It's not something you can enforce, because corruption would ruin it eventually. I'm not some luddite, and I think people should be able to get rich enough to never have to worry about money if they contribute something that increases overall global productivity by a lot more than the average person, so capitalism is probably the only realistic way to do that. Where it breaks for me is that I don't think anyone should get rewarded to the point where they're a billionaire. For example, when I estimate the human cost of a $500 million yacht, I think of it as 10k+ lifetimes of work and consider it a misuse of resources. That labor should be directed somewhere else and no one should have enough money and power to "waste" it on something as superfluous as a yacht.

What I really meant is that we need changes in attitudes and education. Instead of viewing the current tech leaders as geniuses for aggregating users, limiting (or eliminating) choice, and charging for access to users, which is a cost that gets passed to those users, we need the average person to understand the damage done in terms of lost innovation, increased prices, and wasted labor.

I think teaching financial literacy in schools would help a lot. People just don't understand how badly they're getting ripped off IMO.

babypuncher• 2 years ago

There's a middle ground, where we use free market effects to encourage competition and bolster progress, while not letting monopolies and other progress-stifling mechanisms take form.

You can have all the benefits of a free market without also letting someone like Google have the kind of control over the internet that they do.

crabbone2• 2 years ago

capitalist would argue that markets are you best shot at doing this.

That's predicated on anti-monopoly laws and other kinds of fairness requirements that are imposed on the market from the outside.

Markets, or capitalism, if you want, work best when there are some things in infinite supply (eg. consumers), when marketing is honest... whoops! And so on.

Regulations is what's preventing (free-market) capitalism from imploding. But, of course, it's hard to make good rules, and especially durable rules (because people learn to game the system).

jamilton• 2 years ago

A capitalist would argue that, but they would probably be wrong, given massive long-running "market failures" in improving the planet and average quality of life on global and local scales despite reasonably free markets. Markets being efficient doesn't imply positive outcomes/externalities.

I could definitely see there being trade-offs, like non-market solutions resulting in negatives but still overall improving on those particular axes, but I find it unlikely that markets are the best possible strategy.

FredPret• 2 years ago

You can oust any business leader you want by out-selling them (easier said than done!)

But there is one tiny upshot of the monopolistic platforms - it simplifies the go-to-market strategy for all third parties. Hardware and software vendors had a much easier time simply developing for PC for a good two decades.

Now of course we have the web which makes software development even easier than that, but companies like Nvidia have a harder time dealing with multiple platforms than in a Windows-only world.

andsoitis• 2 years ago

I’d argue that’s a huge upshot. Countless software makers have an easier time making products with more homogeneity in the substrate.

aleph_minus_one3• 2 years ago

I often think the greed and bad acting (ex: platforms as silos) we see from tech leaders that are idolized

Among my acquaintances, I honestly don't know a single person who idolizes these tech leaders. So this statement depends a lot on in which circles you spend your time/life.

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.

By your conscious buying decisions, you decide which business leader you want to reward. It's that easy.

donmcronald4• 2 years ago

Among my acquaintances, I honestly don't know a single person who idolizes these tech leaders.

I've talked to people that think Steve Jobs and Elon Musk are going to be standing beside St. Peter at the pearly gates.

Paul-Craft2• 2 years ago

By your conscious buying decisions, you decide which business leader you want to reward. It's that easy.

Let's not blame consumers for the bad actions of corporations, please.

RcouF1uZ4gsC• 2 years ago

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?

They actually moved humanity forward. Microsoft/Apple/Google have brought computing to the masses. They made computers work for the average person. Whether from the goal to put a PC on every desk, put a powerful portable, usable computer in every pocket, and brought turn-by-turn directions to the masses.

The whole Unix, designing only for the high-priesthood and everybody else be damned philosophy has probably held back computing more than anything Microsoft or Apple or Google did.

donmcronald4• 2 years ago

They actually moved humanity forward.

Yes, that's undeniable, but what I'm saying is I think they did a bad job and that we could be further ahead. As a small example, take the time wasted on developing for IE6 because anti-competitive practices were used to prevent alternatives. What's the opportunity cost in terms of the innovation that could have been achieved if all those development hours went into something more productive than fighting with a crappy product from an entrenched incumbent?

What if Apple would have embraced Java and Oracle hadn't fumbled / abandoned JavaFX? I would argue that developing with JavaFX was a better experience 10 years ago than WebDev is today, at least for line of business apps. Imagine if every technology like that had been allowed to survive (and thrive) everywhere. Developers would have significantly more choice and we'd have much less waste in the context of finagling with the current options to accomplish what we want.

The whole Unix, designing only for the high-priesthood and everybody else be damned philosophy has probably held back computing more than anything Microsoft or Apple or Google did.

For a long time I've thought a lot of open source projects are "good enough" projects sponsored by big tech to ward off competition from high quality alternatives. As some specific examples, BTRFS with the parity patches that were rejected about a decade ago because they didn't match business goals or the WinGet fiasco that killed AppGet. Those are opinions. I'm not accusing them of anything.

ycjspjh• 2 years ago

It's like saying it's possible to get rid of all the predators and parasites in Nature. But for some mysterious reason Nature seems to keep them around.

I think it's more about using them(and their seemingly infinite yet mindless energies) to push things in a certain direction and reigning them in when they push everyone around.

Sort of like domestication of dangerous animals.

Aerbil313• 2 years ago

Holding back the humanity from what? Technological progress?

I encourage every technophile to read Technological Slavery by Dr. David Skrbina: https://ia800300.us.archive.org/21/items/tk-Technological-Sl...

Obscurity43402• 2 years ago

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

brudgers• 2 years ago

Farnham Street is a part of The Economist.

I didn’t write it.

Obscurity43402• 2 years 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• 2 years 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_one3• 2 years 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• 2 years 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• 2 years ago

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

11101010001100• 2 years ago

Does this explain our response to climate change?

29athrowaway• 2 years 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• 2 years 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• 2 years 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 :(

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

Related Stories