Could the tech companies run *everything* better?

Under one view, the major tech companies lucked into some pieces of rapidly scalable software.  They are phenomenal at producing and distributing such software, but otherwise they put on their pants one leg at a time, just like the rest of us.  They are not especially productive at marginal activities beyond their core competencies.

Under the second view, the major tech companies have developed new managerial technologies for hiring, handling, and motivating super-smart employees.  That is the reason why the tech companies have become phenomenal at producing and distributing rapidly scalable software.  But if tech companies turn their attention to other productive activities, they would do very very well.  Alex for instance thinks that Apple ought to buy a university.  Or you might expect that Google’s “scallion fried fish” dish would be especially tasty.  After all, do not smarter people make for better cooks?

Yet a third view starts with the idea of labor scarcity, at least for the very talented folks.  Good, ambitious, non-risk-averse managerial talent is super, super-scarce.  The tech companies have a lot of it — good for them — and they pay for it by producing and distributing readily scalable software.  In that setting, there is usually some slack within the tech company, so if the tech company takes on a new activity, it will excel at it, at least provided it does not try to move beyond the margin allowed by its collected, on-call talent.  Yet if the tech company were to undertake a massive expansion into many non-tech fields, it would be just as talent-constrained as anyone else.

Which are these three views is correct?  What if you had to pick three percentages that sum to one?  How about 30-30-40?

Is there another contending view I am missing?

Addendum: A very important question is at what rate the existence of the tech companies boosts the incentive for individuals to become one of these very talented cogs in the machine of grand productivity.  Training and talent-spotting matters!  And just as tennis players keep on getting better, so can we expect the same from talented, high-cooperation workers, at least as long as the rewards are rising.

Is this actually the variable that determines how much good the big tech companies do for the world as a whole?


It's obviously down to luck in the cases of Google, Facebook, Microsoft etc. One lucky break - then hiring and paying talent well to sustain monopoly power. Google might be the worst company in history in terms of new product development hit rates, so everything would probably be worse run by them.

Google glass

Also, they don't mind forcing people onto their products (for example google+), if people don't start using it voluntarily. They have the size to spend money on projects that fail they have the size and power to force the users of existing products onto their new products, and they have no costumer service at all, so it doesn't help to complain. You can't get in touch with them. You can only write on their product forums, which nobody from google is going to see, unless a million people have written about the same thing.

So, which mapping software company already making money before Google started doing mapping is offering a better and more successful mapping service?

Which machine text translation company in business before Google got into translation is offer a better and more successful service?

Which mean text recognition company in business before Google got into machine translation offers a better and more successful service? Note Xerox and Adobe were doing this before Google.

Which image feature recognition company in business before Google got into it...

Which email company in business before Google got into it offers a better and more successful service?

The places Google beats incumbents is where paying lots of workers a lot of money over years creates user demand.

Paying a lot of workers over years on such things as Google Glass resulted in little user demand. The biggest problem was, is, "whats it good for?"

Google getting into smartphones is a failure, or success? Google did very little new at the start, except giving away a lot of its enhancements to Linux embedded software. Other smartphone makers did better enhancements at the time, but didn't give them away. And a number of smartphone makers had developed better software for their products. E.g., BlackBerry was much better than Google Android for phones.

Google was hardly the best search engine for years, but Google was willing to pay every penny plus some to workers, while the competitors wanted to "create wealth", "liberate value", for instant profit during the boom. In the 90s, I used 4-5 search engines, but mostly AltaVista, which was expanding the data sources that could be searched. AltaVista was the best at searching documents, PDFs, Word, etc

Then Compaq spun off AltaVista and other software to "create wealth".

Is that decision by Compaq's new CEO the "lucky break" Google got? When he was explaining how great he was, he justified buying DEC, a move by the previous CEO, and dismantling half of DEC to cut labor costs, based on the stock value of spun off Altavista Corp at more than a billion. Which was vanished two years later.

Google was a loser at that time, failing to "create value", etc. So, the late 90s winners, Yahoo!, AltaVista, and others, are where, today?

Exactly. Google is successful because they have a monopoly on search, MS on desktop software. If you go just a bit down the line, it is hard to argue that a Dell first line manager is risk tolerant and you wonder why a company like Symantec is even in business. Google will hire a small number of people for useless but high-profile projects like self-driving in order to gain media approval, cherry pick in hiring, pump the stock valuation, and distract regulators from its monopoly.

Tech attracts smart people because it treats workers OK - it pays well, you can work from home, etc. To the extent other fields do this they can gain the benefits of talent, but no government job (other than some sports coaching) pays well and even if you had a free hand universities aren't particularly good at conveying information as you can just buy the textbook. Apple hates freedom of information and their tech docs aren't all that clear at why they makes thing five times more difficult than they need to be and most food tastes OK. Even the successful tech companies aren't particularly well managed, they are mostly managed by former tech workers who got tired due to old age and family obligations.

I guess the value of the third view isn't clear - are we talking about tech talent being better at the business of running a restaurant chain or actually making better tasting food? I suppose tech workers could do things like write SW to predict the best location to expand to or for scheduling shifts - tech also has unrestricted free access to financial capital. I suppose tech could also make new kitchen appliances, for example to cook something on all sides instead of just one at a time. But I don't think the benefit would be all that great. So I go with 78-2-20.

Google has no more of a monopoly on search than Altavista had in the bad old days. Indeed, search engines have very low moats, and if a significantly better (or easier-to-use or better-integrated) engine comes along, the "monopolist" is dead within a few years - as opposed to real monopolies with network effects like the decades-in-building desktop OS ecosystem of MS or the collective action problem of migrating away from Facebook. Google's main line of defense is hiring the smartest search specialists so that no other engine gets a large enough advantage to switch - which is the premise of this post.

As an observer of my sons's Microsoft XBOX use, everything about it is terrible, except for access to the key games like FIFA and Madden. The controller does not have an off switch, you either go through some menus to turn it off, or you have to take out a battery, or, more often, my son just leaves it on to run out of batteries.

Its pure grabbing on to the network externalities in this pretty important domain.

"I can't believe I still have a 'C' drive."

- me, 15 years ago.

The innovation rate seems to be related not only to shear smarts, but also to the market the company is in, specifically, how capital intensive the industry is. I worked with very bright people in the aviation industry. While individually they were quite innovative, the company and market itself was not. There was no huge (profit) upside to dramatic improvements, but the cost of product development and failure was very high.
I think that software related fields allow rapid innovation, as the capital cost of trying new products is low, as is the cost of failure. As the products become more capital intensive, and more difficult to get to market, innovation goes down, no matter how smart the folks involved are.
This is going to show up with products like new space craft, new electric cars, etc. You'll see innovation slowing and overhead costs going up as the products are more widely used. Nobody cares if their cell phone stops working now and then, but one major rocket explosion and you've got an army or government and insurance folks clamoring for new regulations.

TC: "Is there another contending view I am missing?" - yes, that employees, when they sign standard employee agreements assigning all their intellectual property to the employer, are hindered from producing innovative things. I've actually been told many times by talented employees that they are "holding back" from disclosing inventions, until such time that they start their own company and can patent the IP in their own name. Very common.

I have seen some weird things, but in California there is the legal framework for the honorable employee who gives all 9-5 and invents something else after 6pm.

Much of that after 6 work, which would be the subject of legal jeopardy in other states, has been the launch of a new company. Land of the free.

California is more employee friendly than other places, for one thing they narrowly construe all non-compete agreements, but even in the after 6 pm scenario the employee would have at least shop rights in the invention, meaning a cut in the invention. Even Google had to pay Stanford University as I recall for the shop rights to Google's search engine, since it was developed with Stanford U facilities.

I meant after 6 and on own equipment. That hasn't been hard since PCs allowed average programmers to own the means of production. Or now when cloud vendors offer experimental accounts.

But yes, if you use company resources it becomes company work.

Using outside research grant money. DARPA and NSF.

In the 80s, conservatives argued for, and got changes to the law on IP from Federal grants. Their idea was the grantee, the university, typically,would hold onto the IP and license it instead of it going into the public domain with all rights assigned to the Federal government which granted everyone free use in most cases.

They argued two things.

1. The Federal grants could be eliminated because universities would get so much more money from licencing the IP.

2. Research would result in innovative products quickly by creating monopoly control over Federal funded IP, ie, free patent rights prevented them from being put in products because anyone could, so using the IP generated no profits. Ie, with two or more firms using free IP, the highest price would be all-in labor costs. Zero economic, monopoly, profits.

In that argument, the only reason anything is done is for profits. So SpaceX success is due to extremely big profits, higher than most tech companies. Tesla is even more profitable, profits rivaling Google. Google stock price is up less than 3x in five years, Tesla is up 10x.

Note, the Internet IP, especially all the code, was developed before these changes from all IP from Federal grants going into the public domain as free to almost all. (Munitions could not be exported freely, even if freely available in the US.)

DEC engineers replicated the pages rank system in AltaVista- their research papers were public, so recoding and improving it just required paying workers. It was a simple competition between DEC and Google on improving on page ranking software for several years, until Compaq took AltaVista out of that competition and focused on ad revenue instead of better search results. Selling ads requires worse search results..

TC: "Which are these three views is correct?" PhD Economist asks us rank amateurs for the answer? Is that how things are supposed to work?
Seems like the economist should be giving the world the correct answer.

Wisdom of the crowds, bro.

I work in the tech industry and I’d put maybe 80% on option 1. Amazon a notable exception.

I don't work in the tech industry, but I'd certainly agree on Amazon being an exception. I think with the others, what's helped them as well is good decisions at the top (acquisitions in the case of Facebook, bold and successful initiatives in the case of Google 2004-2011) and that probably differentiates them from Yahoo, rather than execution. It's probably not sustainable though.

Amazon is retail. It is a well defined problem with many many solutions. Amazon does what it does well, but ultimately there are lots of people who do as well as they do.

No, they're a cloud services company. Retail is more like a hobby for them.

"The problem was obvious. We didn’t have that infrastructure. So we started building it for our own internal use. Then we realized, 'Whoa, everybody who wants to build web-scale applications is going to need this.' We figured with a little bit of extra work we could make it available to everybody. We’re going to make it anyway—let’s sell it." - Jeff Bezos (

How did Amazon not "[luck] into some pieces of rapidly scalable software" again? Without AWS, Amazon would just be a huge project in subsidizing logistics and ecommerce with investor money.

The skill isn't having, or stumbling upon an idea. The skill is executing on the idea. And Amazon are *incredibly* good at executing on those ideas. AWS is the best and most visible example, but certainly not the only one.

In actually executing, Amazon are better than Google, and incredibly much better than Facebook, Twitter or Snap. If Amazon ran education, they would probably succeed. Google, MS and Apple would possibly succeed, but I'd bet against it (because most new things fail, so betting against it is always the best if the odds are even money). Facebook, Twitter and Snap would definitely fail badly.

I work at Google.

I agree, mostly on option 1.

Options 2 and 3 can't be neglected, though. Google has amazing support staff and internal tools. It's easy to use massive datasets or build scalable infrastructure. Is that going to help run a university? Doubt it, but it makes it easy to build a billion dollar cloud business.

Another point on option 3: Some projects (self-driving car) are deliberately set up as retention devices for engineers from search, etc. It keeps people at Google instead of working for FB. There's also the "halo project" effect, to dangle cool projects in front of new hires, who mostly work on more boring projects.

Enthusiasts for "Agile" ways of working believe that lessons learnt about how to organise the massively complex task of creating computer systems are generally applicable elsewhere - see and

I believe that we are much better at creating web sites than we were even ten years ago, and there are spectacular examples of the relative productivity of different skill levels at this - as was demonstrated by the successes and failures encountered by those producing web sites in support of Obamacare.

I believe that there are indeed ways of working in the non-computer world that would be improved by simple Agile methods such as daily stand-ups (See in particular:
What did I do yesterday that helped the development team meet the sprint goal?
What will I do today to help the development team meet the sprint goal?
Do I see any impediment that prevents me or the development team from meeting the sprint goal?


1) My experience with the introduction of Agile is that it is always massively influenced by the existing culture, so it may not be as ground-breaking as you might think. Cultures that expect a single all-powerful dictator still have one after Agile. Cultures that are mired in multiple layers of bureaucratic monitoring and progress report writing still manage to find a way to mandate those reports after Agile.

2) Some of the progress is due to the introduction of specific tools and to accumulated knowledge about those tools (web application servers, automatic test frameworks, source code repositories, Object-Relational Mapping tools...) and this part of the progress will not generalize beyond the domain of computer software productions.

Methods are important, but so too that when software fails, it does very clearly. This builds both systems and cultures based on correctness.

In software, correctness is easily named (software performs to a rigid definition, often to a specific goal).

Companies that just want more quarterly sales, and can't define a "correct" burrito or customer experience have a harder time applying tech.

Note that the few companies who have scaled from non-tech beginnings (Walmart) used computers to pursue a "correctness" of their own.

When you think about it, Facebook and Twitter got lost when correctness became less clear. It was easy when "everybody can post messages" was all you needed. If you wanted to "optimize for engagement" that was easy too.

But all of a sudden "everybody posts" and "engagement" had a dark side, and there was no easy metric for correctness to fix it all.

Now Facebook and Twitter are as lost as anyone in "but what should we do!"


Only in the imagination of hyper partisan leftists is the biggest challenge facing Google and Facebook policing against posting right leaning content. And, again, only to the hyper partisan leftist do these fringe cases justify a public square monopoly censoring all right leaning media. Before you object to that claim, note that videos of Mr. Hogg failing to give prepared lines were censored as the leftists used fringe conspiracies about Hogg being an actor to censor all information that suggests he is anything but a neutral teen responding to a school shooting. There is a world of difference between conspiracies that the shooting didn’t happen and these are paid actors and the truth which is that the shooting did happen and these teens were hardcore leftists beforehand and are now being guided through media, assisted with who to talk to and what to say, and handed over twitter-created verifies twitter accounts with tens of thousands of followers. The nuance was lost (intentionally) by the left that points to infowars to ban Praeger U.

I wonder if "Thomas" is some kind of performance art project.

How is fake news a real problem for Facebook? This is something I never understood. Does fake news negatively impact earnings or threaten future earnings? Isn’t the whole point of Facebook having news in order to cocoon users in an echo chamber that steals away their attention (and time) in order to sell adds? That is to say that it is essentially the boob tube with user generated content and links to some newspaper articles?

If you were intimately acquainted with how the sausage is made you would build a shelter with 3 years worth of canned food.

They don't appear to be any better than the academy at creating an environment of open intellectual debate:

Tech companies are much, much better than non-tech companies at producing software and using software to run their own businesses. In areas where using software well is critical, tech companies will crush non-tech companies. That is not all areas of the economy. However, as software grows more powerful that may be more of the economy over time.

The key that makes software different is that many common business activities, from arranging a meeting to communicating with customers, can be automated and improved via software. At Google or Facebook, the same high-quality employee base that is building their products is also working to improve internal tools.

Software doing what? I have watched companies implement 'software' solutions that had little to do with their value added in the economy, and they were lucky if they survived.

The successful one have the scale to be able to afford a highly skilled team of programmers and implementation experts who design solutions very specific to their needs. They are experts at process automation, the C or Fortran or Java expertise are tools to accomplish a purpose.

I pick something that resembles 1: Software-first companies have better chances than manager-first companies, because the way traditional large enterprises manage is hilariously ineffective. I do not value Scott Adams very highly as a general social commentator, but modern traditional corporations really are a mix of Dilbert and The Gervais Principle. Companies that do the same things with fewer people, and fewer layers of management, have a clear advantage, and enough application of quality custom software leads to high efficiency, which provides superior leverage when trying to get anything done. That said, the worst parts of traditional management are doing their best at getting into large SV giants, so I suspect that they will slow down some in this decade.

I do not believe in 2 at all, as their hiring technology comes down to just paying double. Their early hiring strategies, where they were really innovative, product wise, came down to 'hire from elite universities'. Within the software industry, you will find much derision for the google interview, for instance. Amazon will hire anyone with a pulse, but voluntary attrition is high: Not quite Bridgewater high, but higher than you'd want. They don't do so great at hiring senior people, because their culture is garbage.

High quality management is incredibly scarce, even among the tech companies, and they aren't great at identifying it. Ultimately, they get some of it, just because they pay double, but you will be hard pressed to claim that it's great. If there is a place where I've seen major differences is actually in executives: Your typical traditional company C level exec has a very different route there than they do in tech companies, and in successful tech companies, they tend to know what they are talking about, and can go deep into their company: They are modeled after Gates. In comparison, the route to the C levels in traditional companies involves being good at the upward looking side of management, and not at actually managing anyone, or understanding their business: 6 month reorgs mean that you ultimately don't care about any given project or business. This leads to execs that, at best, know about finance, but that are completely detached of what makes their operations successful or unsuccessful, so they might as well operate at random.

Regarding the addendum, the reason tech company employees are getting better is not really the giant companies, but the gigantic experiment that are all the little ones around them. A place like Google moves at a glacial pace compared to many of their neighbors, but there's enough exchange of talent between the large and the small companies, and there are deep enough networks to make sure the best practices are copied. It's harder for the giants to become semi-irrelevant than it was for Microsoft do so in the late 90s and the entire 00s. If anything, the big news is that Microsoft, which was starting to take the path of IBM, is coming back from its Ballmer nap.

This connection between the Googles and Facebooks of the world and the small and mid sized startups around them is severely underrated: This is what wouldn't be doable in many states with strong Non-Compete regulations. I can only imagine how much faster this all would move if we threw software patents down the toilet.

Do other industries in California benefit from the non-compete agreements?

Another contending view: software happens to be a lightly regulated industry, allowing much experimentation. (Think of all the software/apps that never caught on rather than the big commercial successes.) It's also easy to install/update, providing an effective selection mechanism for software that is commercially valuable to be rapidly adopted. Those two factors allowed rapidly scalable software to emerge. Remember, systems as complex as the human body can emerge from random mutation and natural selection --- no intelligent design necessary. The keys are rapid mutation and effective selection.

If applying tech to other areas can somehow break barriers that inhibit mutation and selection, then tech companies can do very well, but not because of their management skill (intelligent design skill) or even because of tech per se. For example, Uber applied tech to transportation that satisfied the same need as taxis but looked different enough that it did not suffer from the same barriers that inhibited evolution. That allows many drivers and riders to rapidly respond to changing market conditions. Amazon can improve healthcare delivery if it applies enough tech such that the resulting platform allows for products that look different enough from health insurance that (1) they won't be subject to essential health benefits and community-rating requirements but (2) they will still not generate taxable income when provided by employers as an employee benefit. The second element is only needed to allow the new products to compete against traditional health insurance.

'After all, do not smarter people make for better cooks?'

Who knew that strip malls were such a draw for smarter cooks?

What have Google and Facebook done besides coast on their network effect monopolies? Have they actually come out with any new major products?

Google has... Google docs, Google analytics, Android/Pixel/Fi, Chrome, and many more...

I'm familiar with those. None of those were particularly new or innovative. And they mostly latch onto Goggle's network effect monopoly.

"Google docs, Google analytics, Android/Pixel/Fi, Chrome"

Docs, Analytics and Chrome are all pure garbage that made the world worse place. Android and Pixel were bought, not made internally.

Tyler I have known many, many "off the chart" smart people.

Typically, one wants them to have the humility to do good in their local neighborhood. That generally does not happen. Which is why we are where we are.

My best guess is that the "tech companies" where the employees are waking up every morning, rejoicing at the possibility that they will run everything better, are little better (the tech companies) than sad dangerous locations where those who want to ruin this world, in the name of their understandable but wrong idolization of their incremental (and that does not mean much unless it is applied to real concern for real people, one at a time, and in the context we are talking about it almost never is) technical skills, are scheming for a world where people like them prosper and everyone else (not that they remember this) suffers. This is 2018 and there are way too many such sad dangerous locations. They look nice in the morning fog and the afternoon brightness, of course, and under the California moonlight, but they are sad and dangerous for the rest of us, nevertheless.

For God's sake, consider the fact that "smart people" are generally not kind and generally do not care about people who are not smart, and want to bring about a future where only the people who remind them of themselves succeed.

You know I am not wrong.

assume this is not 2018 but 2028 or some other future date. I remember.

The whole view that tech is brilliant people innovating is Google marketing like Enron did. One can concede Google's enormous innovation in search and massive infrastructure (a spinoff from search), and its exploitation of auction theory, but the rest is pants on and off. In particular, Google's promotion of standards and its revival of dying programming languages has led to monstrosities like CSS (the latest Definitive Guide is 1057 pages long), HTML5, and that continuing plague called JavaScript. Back then they said they can't break the web, but it wasn't worth saving, except to make them the monopoly who could keep it going.

FB and Amazon are different sorts of firms. FB is on a lean operating model with good design akin to Plenty of Fish way back, leveraging digital advantages and not reinventing wheels, and Amazon is a scientific method firm running on lines that Toyota started decades earlier. I don't know how one can copy Apple. If Toyota's model was transferrable, then maybe Amazon's is too.

I would say 80, -20, 40

Which is to say tech companies are actually *worse* at non-core activities. I don't really know the reason - something about success making you arrogant and self-righteous maybe - but I work directly with Apple and Google as a non-software supplier. What I see is that they are overstaffed and discombobulated and much less efficient at getting stuff done, and they don't get better results in the end.

Lots of low MP workers producing social and psychological capitol for managers, supported by hoardes of H1b indentured servants to do the actual work.

GE has been the real life experiment in the combination of tech (i.e., software) and industry (i.e., hardware) and its promise. Boy, was I excited that the boy wonders could produce something besides social media, ridesharing, digital advertising, and the software for hedge funds. Alas, the GE experiment failed. Why? Was it because, as Cowen suggests, GE was talent-constrained? Or did the promise of the combination of tech and industry exceed the reality? Or is the hype of tech just that, hype?

Thank God we kept that little, 20+ year-old fridge we had bought the kids for college.

I'm living the GE failure.

Once, you'd buy a fridge and forget about it for 20 years.

Now, we bought (it's three-years-old) a GE fridge (uncharacteristically the wife bought the extended warrantee) that don't work each week after the GE tech "fixed" it four times, and counting. There is one coming this Wednesday - fifth time in two months.

I have lived by the Wall Street adage "it's not what you buy. It's what you pay." I wouldn't buy GE stock at $2.00 per. No, wait . . .

In the last 25 years, the model of controlling a thing, no matter how trivial, by embedding a processor, some sensors, and some actuators, then doing all the tricky bits in software, has displaced everything else. The down side of that is the expected lifetime is much shorter. The last two major appliances of mine that died, died for the same reason -- failed capacitors in the power supply for the electronics. The expected life for such consumer-grade capacitors is about six years. If you get ten, you're doing really well.

Truth. But, three years' useful life is absurd.

Maybe the extended warranty makes sense.

"failed capacitors" I had the same thing in an out of warranty refrigerator. New board was about $400 but I found out online that it's almost always the capacitor. Very cheap fix.

Low quality capacitors are the new planned obsolescence. It's pure genius that a $0.45 part in a $500 appliance can be counted on to quickly fail, leading to the required replacement of the appliance.

GE appliances are the worst. My GE water heater failed after about three years. I fixed it myself after much effort and several days of cold showers. A bolt holding the igniter in place had worked loose. My mother's GE refrigerator has been making increasingly loud noises for more than a year, but continues to function. Because she's over 90, we're hoping that it lasts long enough to cover the remainder of her life. It's about 15 years old, and every other refrigerator we've had has lasted much longer. A knife accident during defrosting killed my refrigerator, which was 35 years old, and its replacement is a small refrigerator originally bought for my sister in 1968. I will never buy another GE appliance for the rest of my life.

There might be hope for GE. Haier acquired their appliance division a year or so ago. My little Haier is going on 20 years without a hiccup. New, it cost maybe 1/5th of what you'd pay for a GE service call.
The company even has an appealing origin story.

Immelt took a huge, but lean(er) and smart company and got woke. He took the company to where its major focus was extraction of federal dollars, and ran that to its natural end. All the talent is gone.

But, but. Surely he did something to earn all that money he was paid, not to mention justifying the spare plane that followed him around, lest he have to fly somewhere commercial. No?

I'd vote No too. I still feel it's nobody business but the stockholders' what the CEO makes. Immelt really took them for a ride though.

Nice sunday speculation but, Computer says no.

The Gates Foundation provides a natural experiment.

So far, where the Foundation has partnered with other, traditional foundations, to help supercharge their previously exusting programs, they've had some success.

The Foundation's own initiatives, however, have been failures. All of them.

So 1, very much.

Cowen often blogs about mood affiliation in a political context, but nowhere is there mood affiliation quite like the mood affiliation with tech. Consider autonomous cars. Most people envision autonomous cars racing along city and country roads at 60 or 70 mph while the occupants devote their time to staring at their autonomous smart phones. When I remind people of the Google engineer's gaffe, that autonomous cars will be limited to 30 to 35 mph, people deny the engineer ever said it, or if he did, he didn't know what he was talking about, or even if he did know what he was talking about, that was then and tech will solve the equation. Sure, autonomous cars racing around at 60 to 70 mph while sharing the road with teenagers driving non-autonomous cars at 60 to 70 mph while texting their friends. What could go wrong? Some in Congress and in state legislatures are attempting to solve the equation by exempting the participants (i.e., the makers or the autonomous cars and the software that makes them autonomous) from any liability for the mayhem. Who will want to ride around in an autonomous car after viewing pictures of the casualties. Not I. Rational people would solve the equation with a separate right of way for autonomous cars, Elon Musk proposing the construction of underground roads for autonomous cars. Of course, we used to call such transportation "subways". Autonomous cars, subways, whatever one wishes to call it, it's transit, public transit when the public pays for it. And you know who is going to pay for it.

People still get into cars despite the well known mayhem they are known to produce. It's not news because we're used to it.

Like when they had all the hype about Teslas catching fire from the lion batteries. As if gasoline cars aren't sitting on 15 gallon tanks of, well, gasoline.

It's what we have now that is unsafe.

The good thing about getting a little older (59), and reading a lot, and not using the memory hole even when Party commands it, is that you have seen everything before. So I remember the conglomerates of the 60s, with acquisitions generally fueled by growth stock from a successful core business, together with some financial engineering, and the conceit that good managers (i.e., ours) can run any business, exploiting synergies as they do so. By 1980, that was all gone.

Well, if it hadn't been for that pesky fake news peddler Jsack Anderson, ITT would probably be a household name still.

It's possible that these firms are successful through a combination of software skill and management systems, but only in secific conditions. When a) technology can solve a large problem to unlock considerable value, b) the payoff comes relatively quickly, and c) there is considerable financial leverage as the company grows, the model works. A key dynamic is that technology-centric solutions are correlated with financial leverage. These conditions allow a very high failure rate and initial cash burn to support a positive return on capital. When these conditions break down, the time to positive returns extends and there is a negative return in capital. Take healthcare, where the silicon valley model has not been very successful. In lower-return situations, although some management techniques can have very positive impacts (e.g. agile), other management systems focused on risk management are more critical to success and the high tech approach fails.

While I don't buy the talent scarcity argument as it's not clear that these firms as efficient at surfacing talent, there is a gulf opening regarding the ability to use data for managerial decisions in a sophisticated way. The tech and finance industries have sucked up much of the information system development, data analytics, and management talent that can effectively use data, with little cross-pollination of people. This is a real issue for incumbents as sophisticated data analysis generates increasing returns, but not likely to solely predict success or failure.

software in the healthcare industry is lousy. Have you ever been to the Medicare CMS site? I use it and curse software developers every day. Could it be that industries where govt bidding is involved are constrained? Govt contractors and private industry can't compete with the big tech companies hiring practices? So I think the big tech companies could do a better job but they don't appear to want to compete in some industries-they want to "swing for the fences instead" looking for the Next Big Thing.

Medical records needs to be sold as a service. Develop it as a whole package and sell it or the govt will want to get involved with every decision on how you build it. "What kind of backend?" Why do you care? It's FIPS compliant.

I used to be on a team that built software for the Department of Health and Human Services. Not public facing stuff, but internal, security websites and tools.

To answer your questions:

Q: Could it be that industries where govt [sic] bidding is involved are constrained?

A: Yes. Government bidding certainly constraints the process. Having a standardized, top down, slow/low feedback software design process makes for awful software. Really, this is a requirements and procurement problem though. If you are a good manager and know what you want out of contractors, you could get great results. The lack of management skill isn't so much within the contracting firms, but government itself.

Q: Govt contractors and private industry can’t compete with the big tech companies hiring practices?

A:Eh. Probably true, but more important are the incentives produced by different business models. Big tech tends to go straight to the user experience (Google:search, Amazon:logistics, Facebook:social status). Gov contractors are at least fives steps removed from a citizen using the CMS website (procurement <-- client <-- middle management <-- customer support <-- citizen). When I consulted to a Fortune 500 oil company it wasn't any better. It really opened my eyes to the idea that private markets are not really more efficient than public systems. In both sectors, smaller and closer to the user is what produces better results.

Q: So I think the big tech companies could do a better job but they don’t appear to want to compete in some industries-they want to “swing for the fences instead” looking for the Next Big Thing.
A: Big tech companies that do serve public sector (Palantir, Amazon, Lockheed, General Dynamics, Microsoft, IBM, etc.) become masters of bureaucratic manipulation. The incentives being what they are do not make their services all that much better. Check out this article from 2015 ( on how Silicon Valley types saved the CMS website. Haha! The article tries to extoll the heroic ex-techs, but there isn't much to extoll. They failed their main goal. They were hamstrung by regulations. The managers and schedulers slowed any pivots they attempted to make. Gov chewed them up and spit them out.

Anybody who thinks that injecting some silicon valley managerial smarts into a big bank or the public sector can magically make it run better needs to rewatch The Wire. I find it is often the academics, like Professor Cowen, who have the greatest faith in the managing class. Anybody who has worked in corporate/government America knows better, especially those who have been managers themselves.

Anybody who thinks that injecting some silicon valley managerial smarts into a big bank or the public sector can magically make it run better needs to rewatch The Wire.

I'm currently a Technology Manager at one of the largest banks and a frequent poster here, although posting Anon2 in order to not implicate my employer. Previously, I've consulted for many Fortune 500 companies and government agencies. I've also started an ISP and run Ecommerce sites for some of the most well-known brands in the world, including in a dotcom era startup. So I'm pretty familiar with "silicon valley managerial smarts" compared to large financial institutions.

At least 80% of the technology resources in a heavily regulated industry like banking are being spent on government compliance of two types. First, there's the basic regulatory compliance, which defines WHAT you're supposed to be doing in addition to trying to serve customers. Second, there is the security and risk compliance, which defines HOW you must do everything. We have a large directly assigned team of regulators from multiple agencies. It's all based on either the latest "best practices" document a government bureaucrat released, or else the regulators whim that month. They have to regularly show their bosses that they're actively regulating their assigned bank, so they have a lot of whims every month.

As an example, the departments which just patch devices according to the Federal government's required monthly schedule are 4x as large as the departments who are supposed to be keeping them working. 70% of the "keep things working" departments' work is driven by the breakage caused by the patching departments every day. This is an environment with hundreds of thousands of devices in 50K+ locations around the world. We have multiple "single sign-on" technologies in use. Why? Because the regulators mandated different things for different environments at different times. Stuff like that goes on and on...

So sure, "silicon valley managerial smarts" would be a big improvement, but only because we're currently using "government regulators who don't care much about the customer/business side of things" as the people who have the last say in regard to our technology practices. If those went away, the current management could function a lot better, with no "silicon valley managerial smarts" required.


Thanks for sharing, Anon2. If you're in the New York Area, I would love to buy you a cup of coffee and learn more about your perspective. Just send me an email at

These companies tend to have clusters of smart techies and product managers who create super-innovative products. If you have a world changing product, how much management perfection do you really need? And outsiders don't see the potentially world changing products that didn't make it into the world because of bad decisions. Note that when one of these companies tries to compete with a top quality product already in the market, they often fail. Google+, Facebook email, Apple Maps.

Very strange post. Software businesses are unique in many ways .. most notably the business models in which initial R&D is huge, marginal cost is low, and network effects are often substantial. Also the workforce needs to be an unusual type who can be convinced to sacrifice having families and personal lives in order to "change the world" (i.e. produce some junk to make the VCs rich cf. Hubspot) until their careers end at age 40, so geeky young males and immigrants on exploitative visa arrangements predominate. There was already a wave of automotive tech companies that failed to outcompete the Detroit supply chain.

I agree that network effects are key, but I think that argues for 3. Good, ambitious, non-risk-averse managerial talent is super, super-scarce .. and that will lead them to high return ventures. The largest network effects (everyone uses X) provide that return.

Who were the automotive tech companies that failed? When I think of innovation in automobiles, I tend to think of Toyota.

I'm a loyal MR reader. I love the question today about tech firms talent management technology. I am a strategist and researcher in a global talent attraction marketing firms (we use tech and creative solutions). I do a lot of research in firms on something like this question and I think you are right: this is the variable that makes the most difference to how much good tech can do for the world. But that may also be why it cannot be replicated very much in other industries. Talent (and this sort of talent especially) and talent culture does not always transfer well en masse, mostly for cultural reasons, but some structural as well.

1. Google is unique because it is doing a narrow set of things with highly specialised skills. Most other businesses are not. When the vast majority of people in a business are coders or engineers then you can create a niche talent management technology that is highly efficient...

2. Especially if there is a corresponding mono-culture. Jaron Lanier talks about this persuasively. And the recent stories on MR attest to it as well. But it's deeper than that. The motivations and requirements talent have are quite specific. Money seems to be a big driving factor. Some of this may not translate easily to other industries.

2. (a) Perhaps the best example of that is quality of work. The average marginal product of most academics (to take the Apple university example) is unlikely to be as exciting to work on as that of most tech workers; and it is unlikely to contribute to something as important or compelling as tech workers.

2. (b) On that point, tech emerged very recently in a few clusters and has developed over two generations into the current situation. Hiring bankers or academics or lawyers inevitably comes with more history and mess. Opening one university will probably work well and provide innovation. But how scaleable is that really? Ray Dalio says the extraordinary transparency he uses would only work for about 30% of people. I suspect it is less than that. Doesn't tech face a similar problem: their hiring and management skills are refined for a narrow segment of talent. Other elite talent is more varied or more entrenched in other similarly narrow cultures. Are tech's managerial technologies broad enough to break those equilibrium traps?

2. (c) How productive would the average person be on the Netflix benefits package? That works because they ditch non elites. If that was a faesible strategy for other industries they could do it, or more of it...

3. Is there also a point here that the narrowness has brought incredibly high returns from talent to the business - such returns are not available elsewhere, partly due to the talent that is already trapped in the low equilibrium, but also partly due to them not being tech firms? Google puts a lot of time into this, using on the ground management as well as HR. Is that faesible at anything like the same level outside tech? I guess the margins are few where that would happen well for an extended period of time.

4. Does anyone think the managerial technologies they have developed could honestly apply to women? I suspect they don't really know much about women at work.

5. Tenure. Anecdotally, you hear that turnover in tech is too high. 18 month average stints. Could you build an academic department of GMU econ quality with anything like that? The intensity of tech creates that turnover - and it is a good thing - people move and are productive - competition is a sign of a healthy market for the best skills matched to the best work. But most other industries, while they need more turnover, can't operate like that. The return from talent is too low to justify it. Conversely, I suspect that with drivers and so forth - jobs that put actual limits on front line output - it could work if they could learn from tech how to reduce turnover times. Again, that may be possible, but I think the info required may be available already...

6. ...a lot of what these companies do is known from books, blogs etc. How Google Works is pretty popular and summarised in slides everywhere. Interview techniques are replicate endlessly. Where are the gains? There is probably some resistance from HR (maybe) but there ought to be enough leaders interested in this to make it work at some margins already. I guess the required level of involvement from management is the blocker. Have we seen the gains we can get from this without major investment already? If so, does that mean the investment required is too high relative to the limited margins it may work at?

7. The management issue is a separate but related problem. Low quality management is at the root of a lot of performance issues. Can tech solve that, once you factor in the weirdness of their culture? Maybe in some places, but again, I doubt it scales well if at all.

8. This one is a big speculation. I think there is a smart-person's Dunning-Kruger effect. Saying smart people make better cooks is too much of a generalisation. Some of them do. Some do not. A lot of smart people assume they are smart outside their zone. All professionals think they understand politics; hardly any do. Because they are smart they are blind to this limitation. They likely understand it better than many, but really not too well. 'I'm smart at tech so my opinions on X will also be smart' is a potential major flaw with this sort of thing. The entrepreneurs are probably immune from this - see their success at branching out into other fields. Any further down the chain and I suspect the effect grows, perhaps rapidly.

9. Tech gets highly motivated, skilled people turn up who are self taught. Their management technology is therefore designed for a type of talent that doesn't exist in the same way elsewhere. Forms may have a tranche of these people, but how differently can they manage that tranche before it causes other problems?

Overall, I think this means there are gains to be had, but not too much low hanging fruit and some sticky-ness from people that tech may not have the empathy or skill set to deal with. If they did, we might have learnt more from them already.

Very good points, but I would ask how do you know that a university is "correct?"

Can it ever be as simple as running a new software revision against a regression suite?

4 especially is an excellent point.

Yes, and they are finding this out the hard way. I'm sure that there are other industries where the inverse is also true.

a 4th contention is that tech companies are much better at creating ways to internalize spillovers and therefore they will do much better if they go on to business where such spillovers exits such as logistics. Also there are potential spillovers that can generated by mean of new technologies and tech companies have the capabilities to do so

Smart people can also be victims to Groupthink… and so these outfits might be lacking some of that needed alt-wisdom non-geeks can provide.

Here's the thing: Most of the government and the older companies are organized around principles of 'control' and 'information scarcity.' They can't develop methods to exploit information effectively because it breaks their conceptions around control of people and processes.

In many, if not almost all, cases, if the mission is correct and the underlying assumptions that dictate it are fundamentally correct, then the benefits of unleashing information are enormous.
The benefits of unleashing people are also enormous. The costs of control are crippling and asymmetric.

The cases in which tech companies will stumble/fall is where they try to do _the wrong thing_ very effectively. If what they are trying to do the theoretically impossible or destructive...
the difficulty should be obvious.

In many, if not almost all, cases, if the mission is correct and the underlying assumptions that dictate it are fundamentally correct, then the benefits of unleashing information are enormous. The benefits of unleashing people are also enormous. The costs of control are crippling and asymmetric.

That's what I am saying. The further you get from narrow problems with obviously correct answers, the more you bog down in endless and wasteful discussion of what is correct.

Or perhaps skill in managing one type of business is not easily transferable to another.

Look at oil companies expanding into other areas in the late 1970's. Even their move into the highly related area of mining failed. As did all of their alternative energy projects.

How about the tech greats at AOL moving into content generation.

Now another tech great can't mass produce a car.

No, tech companies can't run everything better. And neither can anyone else.

Tyler has missed one of the bigger issues as to why some tech companies did not succeed. Consider how Word Perfect and Lotus both lost out to Microsoft in the early days to MSFT Word and Excel which became the default applications in those arenas. I agree with others that Amazon is perhaps the only major tech firm that figured out a business from the beginning and built it into something that is strong. Apple is a great industrial design company that took a bunch of stuff from others (Xerox PARC) and made some pretty things that everyone wants even though they have to pay a price. Google developed the best search algorithm and knocked Alta Vista and others out of the game. Microsoft cleverly morphed from an OS provider to a company whose applications are the default in the corporate world.


Isn't there an entire collection of science fiction that warns us that tech companies cannot be trusted? Tron, Robocop, The Matrix, Minority Report, Terminator.

I prefer democratically elected villains.

You want it? You got it!

Look at our choice in Nov., 2016.

Congratulations, deranged Scandinavian. First (and only) person to bring Trump into this discussion.

Science fiction isn't actually real.

> And just as tennis players keep on getting better, so can we expect the same from talented, high-cooperation workers, at least as long as the rewards are rising.

Hahaha. Absolute cluelessness.

These supposedly mega-efficient companies that you are talking about work with a pyramid-like structure. At the bottom are the coders (from newbies to people with 5-10 years of experience), above that are the team leaders and architects. Above that are some senior managers, senior architects, and product managers. Generally it's up or out. If you have 15+ years of experience and are not a manager there is little use for you because you ruin the fun of the excitable newbies and give them ideas about having a life outside the office. Also: a certain percent of the newbies are undiscovered geniuses who can be exploited for a few years at wages below what they are worth in the market (especially if the newbies are H1-B). Also you make a higher salary even though the latest frameworks are not something that you mastered in college.

So no. These "talented workers" are not some kind of national treasure, but an expendable commodity with a shelf-life of 10-15 years.

> And just as tennis players keep on getting better,

Maybe this is actually a Straussian indicator

You can accuse tech companies of a great many things, but being effective at hiring would not be on the list. All companies like that have as founding products "lead balloons" that flew - it was nearly impossible to predict that they'd be cash cows.

Previous generations of tech products - desktop publishing., desktop computing which was not subject to the IT priesthood, spreadsheets - all had obvious and immediate advantages. Since then, not so much. I have a running argument with younger people who claim "oh, I do everything on my phone" and the produce of said argument is all the things they *don't* do on their phone.

Jaron Lanier is right. Most computing paradigms are just myths.

Having consulted a bit on some med-tech things, I would say that one of the huge things in play for software companies is that their product area, at least historically, was one of the least regulated in the economy. Consider Google, their clutch application was an efficient sorting algorithm for searches. At the time, no legislation existed which created liability for Google if the algorithm did wonky things (like say make it impossible to get over your past). Similarly with Facebook, they were free to develop their core product without having to ask if it would lead to increased suicides, facilitate stalking, or enable identity theft. By and large the tech companies just did their thing without having to be proactive about anything they did not believe was important.

One of the apps I worked on involved medical devices and allowing docs to get access to them remotely. My employer completely understood how to secure the data stream between doc and device, it was the sort of problem they anticipate and have experience solving. What they had no experience with was HIPAA. If data is medical in nature, you cannot just take the most optimal route to secure it, instead it has to go through a number of legal requirements about who can see it when. Worse, you also need to have the data get ported in EMR software. These are byzantine nightmares that play terribly with each other, let alone with third party software. A very nice idea that would have improved a lot of patient's lives and saved more than few wound up DOA because it is harder to solve problems when you lack carte blanche on methods.

This makes me suspect that tech is going to be less productive whenever they have to play through the same regulatory thicket as other business. Tesla, for instance, has been terrible at getting cars out in mass quantity. The Model 3 has been delayed and with literally billions in revenue waiting it does not appear that Tesla will hit production numbers vastly better than traditional car companies.

Relatedly, tech can get away with terrible optics on its hires: the young (without many dependents), the White/Asian, and the male. In large part they were able to adequately say that they were just hiring who was applying - after all for the 80s and 90s computer geeks were very rarely not two out of three above. Now they have come under fire for "toxic culture" and are being subjected to claims that they are not abiding by the laws and norms of business. To some degree they can still sidestep this by using objective metrics and evaluations based on coding; I suspect that none of these advantages will carry through for non-tech ventures.

I have long suspected that dynamic, lean firms are the natural result of not just new technologies, but also from the opening of a new space where the old legal and social constraints are not enforced. Pretty much all organizations and social combinations eventually have to live with a primary purpose of keeping people content. Dynamism is inherently difficult for people so I suspect that tech is going to be captured by the same pro-contentedness forces and possibly some of their own ossified cultural habits.

You could look at it in waves. Blacksmiths became automobile hackers in the early 1900s. Then came airplane inventors, and so on, until you got to mainframe, pc, and internet waves.

It will happen again. But sure, the Wright Brothers faced no FAA.

+1 for both of you.
So if you industry does not have a mind numbing bureaucracy in place, the government will provide you with one.

Why are you asking about this on a theoretical level? Isn’t there enough empirical data to start answering this question?

60-10-30. If I were to put out a narrative it would be that the tech companies were lucky that 2008 soured a mini-generation of smart people on the finance sector, just about the only other place that will plunk down massive pay packages for fresh college grads with 3.7+ GPAs. Rapid advances in machine learning were also serendipitous, as was Moore's law really working its way into the second half of the chessboard, and of course all this time startups could build what they wanted to build without having to worry about profitability and then the big tech companies could snap up their choice of products. But now Google/Facebook/Microsoft/Amazon/Apple don't need the tippy-top talented people at all, they're basically just bottling Coca-Cola at this point. You can already see an emerging understanding, including within the tech companies themselves, that tech company jobs are rents and ought to be distributed fairly and if you have to go down to the 90th percentile to make it work who cares. GE tried to follow suit, but GE actually makes products so that didn't work.

I agree with the general sentiment that (3) is stronger for Amazon. This is because they constantly seem in search of monsters to destroy. Although their success has been entirely dependent on the fact that they were allowed to take a lot of time to scale and build AWS and so on.

"Under the second view, the major tech companies have developed new managerial technologies for hiring, handling, and motivating super-smart employees. "

Much of this is simply that they do not follow the received wisdom of the rest of the economy, valuing ability rather than credentials. They also seem to be de facto exempt from the prohibition on using IQ tests in hiring. My sense is that they could theoretically run everything better, but in practice they wouldn't, as they would start listening to that received wisdom. If Apple bought a university, it would hire a bazillion administrators, use "holistic"(i.e. corrupt) admissions standards, tolerate blue-hair violence, ect.

If tennis players are getting better, then how come the last 5 winners of Grand Slams have been Federer (35), Nadal (30), Federer (35), Nadal (31), and Federer (36)? If anything, tennis players are getting worse. The next generation can't lay a finger on Fed or Nadal and they are both long past their physical prime. I blame it on new forehand grips that emphasize spin and not accuracy or pace, but whatever the reason, you can't say they're getting better when Federer had more trouble winning at 26 than 36.


First, There's massive survivor bias in the question itself. After all, almost all tech companies fail. Perhaps the options should be:

(1) Is this because almost none of them have a viable business idea, or

(2) They are all managed catastrophically badly, but a few find *such* a lucrative business opportunity that even their management can't destroy it.

The "great company" theory of history is even more flawed than the "great man" theory. It's a drive for an overarching narrative where only a local and academically uninteresting one exists.

Not sure which of the three perspectives this supports.

The cult of personality around Kim Jong Jobs has extended beyond the grave and so his peculiar aesthetic ideals that degrade user functionality find fresh new victims. I'm surprised the Apple Spaceship didn't have all the air pumped out of it to suit some idiotic posthumous spec. The real design genius at Apple lies within their marketing department.

This reminds me of a new building at Stanford, I believe it is the biology building. The windows are all curved green-tinted glass. I was wondering when you buy a building like that, do they give you any spares? If you broke a window, it would be really expensive to have a new one made from scratch. Also, glass (especially tinted glass) gets darker when exposed to the sun. If you had spare windows, you'd probably want to store them on the roof so they would age at the same rate as the other windows.

Mark, Stanford is a nice place, but this is a big world --- I had the test scores to get into Stanford, and probably should have gone. I did not and I am now a lower-middle-class failure, by most 'Stanford alumni' metrics.

But my cat loves me, and my cat is a child of God. Even without Stanford-level academic credentials, I found a cheap apartment with a view of a few acres of woods that would have rendered Tolkien, in his Oxford digs, jealous (or is envious the word I want - I never know). I spent 15 years teaching my dogs to talk and succeeded in that humble task, although I do not think it is all that moral or worth while to explain how I did it. Not to mention achievements in the human world, re descendants. Just saying.

You are correct - they should store the windows in the sunlight, on the roof. (You were the guy, I think,who also explained how to treat wholesale coffee to make it taste its best. You have a gift!) Alternatively, they could, understanding what it is to be a glass window in a solar system like this, merely - well, I will stop there, my good friends include a few patent lawyers - I am no big fan of Picasso but he understood the technique, also (he would have patented if he could, the little mercenary phony).

This side of our glorious Digital Revolution: what are the prevailing US rates of adult literacy, adult sub-literacy, and adult illiteracy? ditto for prevailing US rates of adult numeracy, adult sub-numeracy, and adult innumeracy?

If clever humans could not solve these social, political, economic, and cognitive problems among themselves with preferred strategies of social promotion, post-secondary remediation, wholesale credential and certificate dispensations, bogus accreditation of public school systems, et cetera, would our clever machines (told to deal with clever humans) be any more apt in devising or implementing solutions?

Demonstrably very low: they twice voted Obama president.

"A very important question is at what rate the existence of the tech companies boosts the incentive for individuals to become one of these very talented cogs in the machine of grand productivity."

When I was young, we were called citizens. Now, we are just gray cogs in a faceless death machine...

It's just TC being Straussian again.

So that is what America has become: Straussianism run amok.

Let's face. It, being a global unregulated monopolist rocks.

Tech isn't so much productive and smart as powerful.

Except possibly for Amazon, the leading tech companies couldn't run each other's businesses.

Exhibit 1: Appler car (dead), exhibit 2 Google car (if it works they'll only sell components to car companies)

Living The Dream!

President Trump IS running everything better.

The comments are very revealing about peoples' view of what "tech" can add to industry: a better washing machine, a better web site, on-line education. I linked to an article about GE's vision of the use of "tech" in manufacturing the very large and complex machines GE produces, and, alas, nobody bothered to read it, assuming that it must have related to GE washing machines. Washing machines? It's startling, the low expectations people have for "tech". And, given the low expectations, it's even more startling the extremely high value they place on "tech" companies. My friend can turn his house lights on and off with his smart phone. Why doesn't he just use the light switch? My observation is that "smart" phones aren't all that smart, and that the people who use them get about what they need. I suspect people are impressed with "tech" because of what "tech" has produced: billionaires.

Good point about the scarcity of talent. If we take tech as one end of the competence spectrum and government at the other end we all know that a lot of government could be improved by technological modernization though eventually tech would run up against those same roadblocks put in place by marginally productive people trying to protect their jobs and other constraints. So i doubt that handing government over to say Google would accomplish much.

Everything? Can they even run Whole Foods better?

I haven't heard it's worse, and they cut prices pretty good.

I have a friend who runs a talent evaluation startup in SF. About 6 months ago this person told me that for years, Google has really been the only firm that seriously did internal analytics on hiring (that is, did sophisticated modeling of the information available at hire compared to actual results at the company). Not Apple, not FB, not Amazon---they all relied on heuristics that clearly worked more or less well, but were not more advanced than how smart people hired in decades past. This is apparently now changing.

I know firsthand that IBM worldwide is modeling the employees, an acquaintance PhD psychologist has as main job to detect employees that are underutilized, or likely to receive better offers in the job market, and then intervene. So there is at least one other company. I also know Intel did the same 20 years ago, example:

First annual review of friend of mine: "We have determined that you are underpaid and have an incentive to seek employment elsewhere. Here is a 28% pay raise." The Intel case may not have involved corporate HR.

Having survived the era when Google tried bottoms up management ("smart people will just figure out the right thing to do" after the maybe apocryphal event of Larry threatening to fire all the managers since he didn't perceive them adding value) into the gilded age (those who would otherwise have been bound for Wall Street flooded into product management and upper eng management positions) I can say, (a) full credit for their being willing to take big risks in the early days, (b) the focus on data driven decisions (albeit imperfectly implemented) made a lot of organizational progress in the face of chaos.

Overall, the willingness to use ocd-style consistency and the goal of objectivity give tech companies an edge for running organizations. Some more realpolitik and accommodation of human nature would make them even more effective.

Amazon and Apple are unlike Google and Facebook in being untainted by the give-it-away-put-ads-on-it-later bias. That's a business issue, not a management one.

Obtaining and retaining very smart people requires allowing them to succeed in their goals. This is easy to do in scalable software or electronic hardware where the whole team has the same goals.

When you get into another area like aquaculture of growing fish utilizing high tech instrumentation and controls in automated offshore environments where their talents are very relevant, you are into a business area where dozens of regulatory bureaucrats in dozens of agencies can slow, block, prevent and delay any one of the dozens of "permissions" for years of discussions, lobbying and being nice to idiots. Each bureaucrat can demonstrate his importance by creating a block to getting results. You couldn't keep that very smart staff in these conditions and they may fail.

Observed status of offshore aquaculture: It is rapidly growing around the world at double-digit rates in almost every country except the US. This means that we will import "not so fresh" seafood and frozen product only.

If you believe that “software is eating the world” (which it is), then even View #1 puts the tech companies in a real position of strength.

I remember hearing about how some Apple Executive got put in charge of JC Penny. He started with the idea of eliminating all sales. I also think there was some strange idea about there being almost no one at the cash registers to ring you up or something like that. Anyway I believe they managed to replace him before being driven all the way into bankruptcy.

My wife was there for that - I think it was Larry Johnson?

He had no knowledge or awareness of the customers or the employees.

Neither group of people were ready for the staff to be wandering the floor with wireless devices ringing up purchases. I don’t know how that was supposed to work anyway. For any transaction with more than three or four items you need to have bags and often the machine to take off the anti-theft tags.

The “everyday low price” idea was dead on arrival as JCP had spent years training their customers to be ready and excited for the next sale.

What Johnson really tried to do was trade the existing customer base for a new one over the course of a weekend. Lots of brands put huge effort into attracting new and younger customers but he was much more effective at driving away the existing loyal shoppers than he was at attracting new ones.

Ron Johnson rather

Under one view, the major tech companies lucked into some pieces of rapidly scalable software. They are phenomenal at producing and distributing such software, but otherwise they put on their pants one leg at a time, just like the rest of us. They are not especially productive at marginal activities beyond their core competencies.

I put 70% on Number One and then that helps the companies buy the best talent. The amazing thing about tech is how quickly the companies come to dominate the marketplace with near monopoly market power. So when all else fails, I assume:

The marginal cost of a byte is less than $.01 so they have an endless dropping marginal and average cost curve.

I have been a tech manager for 20 years, ten of those at Google.

Here's what I saw:
Google had very good execution against goals for projects. That was a combination of visionary and exemplary computer scientists and developers.
Google had very good results where the products could be made to work within the sweet spot they were comfortable with: consumer / open source / paid for by ads / ... (less good when they tried to work in more profit-lean, competitive areas)
Google had a warped hiring experience because so many people were pounding down the door to get in -- the "no by default" approach to interviewing, which resulted in many bad experiences for the interviewees was a rational policy. Most organizations aren't that lucky. But, some innovations could only happen in that kind of environment.
Google had innovative management practices which discouraged hierarchy and fiefdom by assigning employees to managers and using hiring committees to select employees, rather than letting managers select or hire themselves (and gain the employee's personal gratitude and loyalty as a result)... -- this changed some as the company got larger and less desirable as a place to work.
Google allowed employees to change projects at will, putting more pressure on managers to serve employees and keep their projects competitive (within the company) so that employees would be motivated to remain and not jump ship -- this also changed some over time and had some unintended consequences.
Google had innovative HR -- hardly a quarter went past without having the employee rating or compensation scheme shifting in some way that required notable adjustment by all managers.
Google was unprepared for the shift that came when large numbers of mid- to upper-range managers entered the company, seemingly having money for a prime motivating factor. At some point the motivations of the employees shifted to be dominated by career concerns (i.e. promotions heading towards more money) rather than overall success of the company, which was more or less assumed.
Generally, you get what you incent...and it's hard to maintain the connection between individual incentives and company performance when a company is big, generous, and successful.

So, IMO, it's worth distinguishing how tech companies operate prior to the point where they are semi-monopoly highly, highly financially lucrative for management class (obviously, in Google's case there was high financial reward at the start, also, but it was counterbalanced by much more risk...), but when they start: employing top talent, using objective, rational measurement, disempowering managers (e.g. discouraging personal loyalty from employees or creation of fiefdoms), and reconsidering problems, including those of organization, from first principles.

Once you dump a ton of money on something it tends to get harder to work with. Would that we all had that as a problem. But, it's worth distinguishing the benefits and problems that come after success is assured and those that come before it.

[Amazon, of course, had a much longer climb and had policies that were overall strategic and highly effective from a business perspective. The Everything Store is a great read. They're not so much a tech company as a company that uses tech very effectively.]

I lean strongly towards *mostly* number 2. Software is different in that it has very low capital costs. It's almost all expense, namely programmer labor. And, top programmers can be incredibly productive -- see . 10x fewer programmers means potentially 100x fewer inter-programmer communications required. My 40 years of experience as a software developer bears that out. Add to that a management culture that embraces experimentation and shuns hierarchy and you have, in my opinion, a recipe for a lot of useful, productive software development. When you start to add capital intensity to the mix, i.e., factories, distribution, retail brick and mortar, risk aversion may begin to take hold, slow down the experimentation and add layers of management. But, many formerly capital intensive activities can now be purchased on an as-needed basis. Think electronics assembly and cloud computing, even distribution and retail if you are willing to pay the "Amazon tax". Perhaps the future successful business model will draw on all sorts of xyz-as-a-service partners to deliver products from a small group of relatively productive and elite designers. As for a university, isn't knowledge one of the first non-rivalrous goods ? Sounds like software to me.

Software companies benefit from making their own business decisions, as opposed to employee unions or regulators.

Software companies benefit from a high proportion of employees that are trained to make major technical decisions. This is a key ingredient for "agile" management.

Programmers have extraordinarily high productivity returns for skill. The difference between a wizard-level programmer and a marginal programmer is about two orders of magnitude - practically all of which manifests in the form of simpler solutions and fewer bugs to fix rather than more lines of code per hour. This means a company that can attract wizard-level programmers for under a million dollars each is at a huge advantage.

No. They couldn't.

From the inside, I don't think we can overstate how correct view 1 is and how shifting out of it really is the act of breaking earth's atmosphere in tech.

> Could the tech companies run *everything* better?

No. 80-10-10. Extremely smart and motivated people? yes. But they are merely racing each other to pick the low hanging fruit in this new frontier. They won't succeed similarly in other spaces. Many edge cases can be ignored in computing (so far... mostly... Facebook/Russia may put this to the test), but cannot be ignored in transportation, construction, health care, etc. Sheer computation only gets you so far. Hardware is a limiting factor (GPUs, LIDAR, rockets, gene sequencers, etc.). And development on that end is much slower. As hardware improves they have the advantage in pushing out software to maximize return on that hardware. But Apple already seems to be running into limits there. Can Google translate its TPUs into real growth? How many Elon Musks can we count on to push the limits in conventional sectors? How many will get as long a leash from investors as has Musk? Probably not many.

Have any of the Google "X" efforts been realized? Self-driving cars, maybe someday, in nice weather. But will they beat the incumbents? Where is the Tesla that was promised to drive itself from LA to NY? Crickets. Cadillac may beat them both, yet.

Look at the pop articles describing the difficulties managing top scientists at Google Verily.

It's one pant leg at a time.

And Google Calico:

80%-0%-20%, because it's mostly about productivity, in two ways. First, there is a sort of multiplier in software, whereby one person's efforts can facilitate the work of many others -- so productivity scales more than in other industries. Second, the variance of individual productivity is huge, with some programmers literally 30 times more productive than others. These are both idiosyncratic to software... the first by nature, the second by happenstance.

Tech companies are not particularly efficient in what they do: they rack in huge profits in their fields because they are Schumpeterian innovators and so reap huge profits before the consolidation of market competition grinds margins down. I don't know why that should be so hard for him to understand: it a pretty trivial result from a basic Schumpeterian growth model.

I believe with technology like Blockchain coming on, there is every possibility that we will witness things working a lot better and in far more profitable way as well. I believe this is the greatest opportunity for online investors, as there is just so much to gain for everyone, but of course, it eventually boils to right choice.

With Cryptos and upcoming ICOs the pick, it is fair to say investment on these are the way ahead. XinFin is one of the hottest deals right now with the epic concept followed by equally cool team and then, it’s running on with 10% bonus for the 3rd round, so in short this is as good as they come.

This point might be related to some of your other points, but I think that it bears mentioning.
Nearly every industry relies on software & technology to operate these days and nearly all of them are terrible at it. They might buy 'packaged' software, but most of it is severly bastardized, poorly implemented and a slave to legacy processes.

In other words, many, if not most larger firms have tremendous technical debt and when faced with decisions around making non-technical staffs' jobs 'easier' (a euphemism for not having to change the way they do their job) and developing good software the tradeoff is made to avoid change more often than not.
I'd imagine (and this is just a theoretical) that software companies have an intuitive understanding of these tradeoffs better than non-software companies and would endure the short term pain of changing process in favor of having good software that is more adaptable, quicker to change and cheaper to support.

Thanks , I've just been searching for info about this topic for a long time and yours is
the greatest I've found out till now. However, what about the bottom line?
Are you positive about the source?

Comments for this post are closed