And yet don't those companies directly employ only around 150,000 workers in total?
Yes, Apple off-shores employment to Foxconn but they are open about their plans to automate their factories and replace all the workers.
This isn't a matter of patience, it's waiting on the never-never.
Productivity gains only drive profits if they provide a sustainable competitive advantage. There is a reason Buffet stays from IT - we improve productivity all the time but he can't figure out where the moat is. Apple (iOS/Hardware engineering), Amazon (Warehouse, Scale), Google (Advertisement market leader), Facebook (Attention) and Microsoft (Office) have some form of sustainable moat. The broad market less so.
He is betting that in some shape or form, they will put that cash to work or issue a nice dividend.
Cloudline is being modest... a stack of a trillion dollars in singles will take you 25% of the way to the Earth's moon.
https://www.wired.com/2011/09/stacking-one-trillion-dollars/
the argument is that when the rest of the economy employs IT at the same level of intensity, the remaining 95% will have similar productivity growth and hence see gains of wealth creation.
we then have to work out how to distribute those gains (hopefully fairly and equitably)
There is a oft quoted parallel with electricity replacing steam in factories- it took a generation or more for factories to stop being clustered around one large steam engine, running each corner of the factory to the speed of one central spindle. the idea is that eventually companies will adapt to use "Digital" rather than adapting digital to fit them.
As a mental sanity check, think about just how, ignoring ML/AI, just how advanced you could rewrite any company you have worked in in past ten years. All the crappy inefficiencies, the pointless spreadsheets. And you have probably worked at the tech end of the economy. is today's IT usage really the best we can do as a society? i doubt it - there seems to be a lot of room for improvement...
So yes - today's limits probably are the best we can do, in this society at least.
Also, just because it's "obvious" that it's possible to re-engineer a company's code base, that doesn't mean it's as easy as it looks.
So it happens quite a bit actually.
"Online shopping came along in the 1990s but retailers struggled to adapt business processes to the internet. They needed to build complementary infrastructure such as fulfillment centers, and, the authors note, customers had to adapt their habits, as well."
The lag between commercialization and widespread adoption is wider than we think.
Customers had adapted in the late 80's to something like Amazon (only problem was waiting for a few minutes after checkout) - but then they adapted to big box stores (made possible by big banks and supply chain software) - which ended up being a huge diversion of resources in the late 90's early 00's. The empty husks of retail are a testament to that.
Internal culture at Google / Amazon, especially during the growth phase, was likely very different than your typical Fortune 500.
It's odd that online grocery shopping is so huge now given Webvan's crash and burn in that space not all that long ago.
It's why i'm positive that there won't be swathes of unemployeed people due to tech. It will happen gradually, as it has always been, and people will have to be shifted into different and more productive employment types. It has always happened this way, and will likely continue to do so.
Anyway, I disagree. I think the sheer speed of the change will create a disaster much sooner than automation quality will make humans obsolete. It took more than a century for the steam engine to replace animal engines around the world, then the electric engine replaced it in half a century. Things are accelerating, and are already much faster than by that time.
What is missing from all these discussions is that human wanting is essential infinite. The more productive you can be at making something, the more it will be taken for granted and a new thing wanted.
Demand is infinite. Work is created to supply it. There is no amount of labour that will ever exhaust it.
All automation does is increase the space of "taken for granted products". The infinite space of potential products/services for human beings to supply still exists.
80% of the UK economy is already services from people to other people.
If we reach the happy situation where robots are providing us with the things we need for a good life almost automatically, why would we continue to work ourselves to the bone for additional goods and services that won't increase our happiness much? Most people would rather just work less.
In fact, our current level of economic output is enough to give every household a comfortable life, even if we worked substantially less.
Unfortunately the system is structured to maximize labour time, keep us economically insecure, and use us all as grist for the capitalist mill, rather than maximizing leisure time and quality of life.
While automation of necessities will make it easier to opt out of the capitalist mill, I would expect that to not be a tradeoff-free choice.
If you improve your factory equipment to produce double the items per employee-hour at a lower unit cost, you likely won't see the demand double in proportion. Most likely you'll just eat into the market share of a competitor. Maybe some new customers will newly be able to afford the product, but once it has "taken for granted" status, even that is impossible.
And we are not increasing the number of these "taken for granted products" very quickly. As far as physical products are concerned, I can only think of the smart phone as a recent example; actually that decreased the number of those products, by obsoleting cameras and dumb phones for most people. Among internet services, advertising is the dominant money-maker, and that's a zero-sum game for eyeballs which only employs a relatively small number of people.
As more and more products and services are provided by machines, fewer and fewer people become necessary to keep things running in the background. Sure, some of the replaced people can try to come up with new things to offer, but most of them won't succeed, and the vast majority will eventually be unable to cost-effectively provide value to other humans at all.
At that point, we'll find out whether it's cheaper to feed them or shoot them.
Drones, of course.
It has been a while that demand stalls. We can see it by observing that energy (time, investment, ...) doesn't go much any more to production means, but to advertisement, marketing, new forms of distribution, and so on. Production is enough in many domains (and there is no room for newcomers, it contracts rather than develops, you just compete to take shares from someone else, not that much to contribute to a booming market), the current problem and aim of companies are to invent ways to get people to consume more and more stuff.
https://www.recode.net/2017/7/24/16022210/alphabet-google-em...
It's pretty simple to determine -- Who's paying the salary? Your example, of a contractor who is employed by Amazon to work on Netflix integrations for Amazon products, would count for Amazon, as the work is paid for by Amazon and is for Amazon's benefit. It wouldn't count for Netflix; we're only talking about employees whose salaries are being directly paid for by a company, not larger second-order ecosystem-type effects. That gets a lot harder. What I'm talking about is exact.