[Note: I originally scheduled this post to be published next week, but I moved it up after listening to the conversation between Mark Zuckerberg, Tyler Cowen, and Patrick Collison. In the transcript, Zuckerberg says
So rather than people moving–inventing a new hyperloop or cars, I tend to think the set of technologies around–whether it’s augmented reality or virtual reality or video presence that just lets people be where they wanna be physically and feel present with other people wherever they need to be to do their job, to connect with the people they care about–that feels to me the better long-term solution.
Those are the thoughts I express and elaborate on below.]
I remember hearing Robert Metcalfe (link goes to Wikipedia) speak about twenty years ago, and when he was asked what he thought was the killer application for the Internet, he said “telepresence.”
I thought of this when I saw the paper on mobility in the United States by Kyle Mangum and Patrick Coate, pointer from Tyler Cowen.
repeat mobility is common. That is, people living in their “home” locations are far less likely to migrate than those away from home.
My train of thought went as follows.
1. I view the paper as showing that many people come to like where they live. The repeat movers are either innately restless or experimenting.
2. When people my age talk about their children’s work lives, a sentence that comes up frequently is, “They let him (her) work remotely.” Of my three daughters, one works in Boston for an organization based in Maryland, one works from home three days a week, and the third probably could continue to work remotely if her husband moves.
3. In fact, a lot of married couples have job opportunities in different cities.
4. Recall that Patrick Collison said that his firm set up a department that he calls “Remote.”
5. As Patrick pointed out in that same conversation with Reid Hoffman, Zoom Meeting is quite a step forward in the videoconferencing arena. I can’t really articulate what makes it better than Skype or Google Hangouts, but it just feels more conference-y.
6. If I were in the venture capital business, I would make a bet that remote work will grow exponentially, and I would assemble a portfolio of companies based on that bet. Will more people wear body cameras? Do small companies need better support for interstate human resource functions? What are the needs of the home-office worker? What sorts of meeting-scheduling systems address the challenges posed by remote work forces?
7. I think that blue-collar work may be an overlooked opportunity for telepresence. Techies talk about telemedicine, but it seems to me that it is much harder to remotely work on someone’s body than it is to do other tasks remotely. So blue-collar telepresence may come first. Professor Daniel Markovitz, author of the Meritocracy Trap (in another conversation I plan to annotate) says that Amazon warehouse workers already are subject to remote monitoring.
–How about tele-sanitation? Bathrooms at places like airports and hospitals have to be cleaned and re-stocked very often, and robots could do that. But the robots might not be able to operate completely independently. A remote operator could help the robot be more adaptable to situations.
–How about tele-chauffer? Even if self-driving cars are not ready for the road, who says that the driver has to be in the car? In the case of truck driving, the number one source of job dissatisfaction is being away from home all the time Telepresence could solve that problem. Perhaps a co-pilot does not have to be on the plane (assuming you want the pilot to be there).
–The highway construction workers who operate machines. Do they need to be there?
–The workers building skyscrapers. Could they operate by managing robots remotely?
8. Think of what Zoom Meeting and other telepresence apps will be able to do when 5G is ubiquitous.
This from 1965 seems to have gone round again:
http://www.quantium.plus.com/ahr/ballantine.htm
>>>
The invention for which Mr- Reeves received the award marked a return to a concept of digital voice transmission which predated Alexander Graham Bell’s discovery by 40 years- Prior to 1876, scientists had tried to “telegraph” music and speech, but failed. Bell’s successful idea was to produce an undulating electric current varying directly as the voice varies.
Recently, Mr, Reeves estimated the future impact of PCM on communications. He foresees that by the year 2000, we may well be a nation of “stay-at-homes,” because of greatly larger populations and a transportation problem that may require executives to handle business by phone, since it will be impracticable for them to commute.
This, plus ever-increasing amounts of literature, will require the establishment of information centers with which each individual has communication. By 2000 A.D., Mr. Reeves suggests, we will be transmitting intellect and information, not transporting bodies and books.
<<<
That is a fantastic quote/link. The Wikipedia page on PCM says Claude Shannon, “the father of information theory”, was co-author of an American patent on PCM in 1946. Audio codecs have an interesting history. Texas Instruments re-invented itself in the 80’s with the introduction of a DSP+Microcontroller SoC that became the basis for most digital feature phones. Skype is remembered for their peer-to-peer technology but equally important was their proprietary low-bandwidth audio codec that worked on generic desktop computers.
Lots more on http://www.alecharleyreeves.com
about this little known British inventor who pre-dated Shannon with a PCM patent. His Equilibrium coder also has a weird echo in our time although it is now called something else. He also has a page on Wikipedia
https://en.wikipedia.org/wiki/Alec_Reeves
If I can have a job done by someone working elsewhere, then I would pick the best-valued person on earth. For the majority of current jobs in rich countries, whoever that is, he or she almost certainly doesn’t live in a rich country.
That is, unless there are (1) legal rules discouraging it (e.g., local license requirements, regulation insisting of hiring locals or citizens, or high tariffs) or (2) some kind of regional skills or knowledge that are too hard for more distant people to pick up, or (3) a requirement for periodic face-to-face interactions for which transportation costs to and from distant places are uneconomical, then I am almost certainly going to ‘outsource’ to a lower-wage country to find the best value.
Which means, so long as large wage-differentials persist and we don’t have unlimited free movement of labor, the improvement in remote-work technology will shift the utilization of the local, more expensive labor force to adjust to a new equilibrium in which most jobs are inherently ‘irremotable’ which, in the nature of things, are services which require close geographic proximity to other people involved in consumption and/or production.
Some people (e.g., Yglesias in The Rent is Too Damn High) call these “face-to-face interaction” jobs, but you never need to see the face of your chef or delivery man to have needed to be very close to where they performed their service. There can be small gaps in time, even for quickly perishable items, but the geographic gaps have to close.
What people envision in terms of a remote-work future is a bunch of people in our own society shifting to working remotely. That might happen for jobs where that can be done some of the time, but not all the time.
The economic logic of the situation means that as remote-work technology continues to improve, it will skip right over people in rich countries and mostly be used to connect rich-country consumers or employers to poor-country employees (which, by the way, means we won’t need open-borders in order to achieve a lot the purported economic benefits that open-borders advocates claim lies untapped.)
Local, affluent people with lucrative occupations in rich countries will find that they simply cannot make any good use of those remote-technologies to lower the burden of commuting, traffic congestion, transportation costs, and high real estate costs near where they work.
That, in turn, will continue to draw workers in to the centers of gravity of the major sectoral and regional hubs (i.e., “Winner Cities), and likely continue to raise the economic value of central real estate.
So, if I were to make bets in the remote-work sector, the bet would take the shape of it unleashing a new wave of outsourcing.
Handle,
That is a good point. The tele-janitors and tele-drivers are likely to include many foreign workers.
We have real life markets that test this out, and almost everything you are presuming isn’t happening. Even in the realm of software coding, where all of the variables seem to have been in place for this to play out as you suggest, it really hasn’t worked out as you suggest.
A substantial amount of outsourcing is taking place, but American coders are making as much money as ever. If you are any good and you want to work, you can. A huge percentage of those American workers are also working full-time online, but still, their jobs are not going away, at least so far.
It isn’t necessary for local legal rules to explicitly discourage international outsourcing. Just plain old normal complexities do the trick quite nicely.
I don’t think the following statements are consistent:
(1) Most software developers would be equally productive if they worked remotely full time from some cheap place, and
(2) There is a shortage of housing in the Bay Area, which is infamous for having one of the highest costs of living in the world.
(3) Big smart silicon valley companies are notorious for setting up shop in the country’s most expensive metro locations, concentrating tens of thousands of employees in one place.
I just had a close relation sign up with Facebook which insists in the contract on the work being performed physically from either the Bay Area or New York City with no telework options, and thus has to offer tons of money to compensate for the significantly higher cost of living.
What makes more sense. To say that these companies with individuals who are the most likely on earth to adopt and exploit cutting edge technologies are walking past $100 bills on the sidewalk, or that they know from experience that, like it or not, the most productive development environments are ones in which team members work in close, physical proximity with each other, and with managers and subordinates?
Handle, I think the three statements you list are consistent when “(1) Most software developers” is replaced with “(1) Many software developers”.
I wrote a (too long) anecdote earlier today that is stuck in moderation. You treat the local vs. remote decision as mutually exclusive when in reality the best strategy/tactic is a combination of the two.
OK, but that really wasn’t your argument before. It isn’t necessary for Facebook to aggregate talent pools like that. They could get an equivalent coder at 1/4th the price elsewhere, and they could make that work. They prefer it otherwise, and all of the tiny conveniences and cross-pollination that it provides.
It’s also a matter of fact that as RAD says below that Facebook and most of the American software industries do employ a combination of strategies.
My point is that there are few jobs that fit as cleanly into a telepresence model as cleanly as coding does. And yet, the value proposition remains stubbornly complex. I doubt it will ever be much different.
Our jobs may get shaved away by automation, but it isn’t so easy to substitute people from a different culture and regulatory regime, even for simple work. It gets done, but its no slam dunk.
I am skeptical. You would never have a professional sports coach work remotely, and the reasons aren’t simply because sports involves physical activity. If you want to solve problems, you want people to very regularly be interacting face to face.
That isn’t to say that remote work won’t increase, but if you see a lot of remote work being done in a company, that is in my opinion a sign either that those workers are getting paid to dig ditches, or that their work is so specialized that it gains very little from relationships with coworkers. In either case, remote work isn’t great for your career.
+1
Automation and robots are coming but at much slower pace than expected. Just think of driverless cars felt closer to the market in 2012/2013 than they do today. My guess is the programming the last 20% of cars decisions ends up being 80% of the work here. And a lot of service jobs you need robots to do 98% of the work because you can’t find replacement workers when the robots break down.
Look at the usage of automated cashiers at grocery stores, order takers at McDonald’s and the credit card purchasing at Darden restaurants. This stuff has been around since 2007ish and yet only the last couple years are you seeing the mass rollout of all of them. And the rollout
1) The rollout of automation in services also needs customer usage so it takes a few years for customers to use. I remember how much automated phone systems were hated in the mid-1990s but consumer got used to them the next several years.
2) The automated cashiers are still focused on customers with few items and really only saves 1 or 2 workers per store. (These systems often have an employee watching customers here but customers like the shorter lines as there are 4 – 6 of these.)
I work remotely 24/7/365. Part of this has to do with it being a very small industry with not many people in the specialty and it not being geographically concentrated. Also, while not all the work requires a licensee, the end work product of our department is something the government requires to be certified by someone with a license.
I think Handle is right about the fundamental issue. When people think telecommute they think:
1) I don’t have to move to New City for New Job.
2) I don’t have to pay sky high real estate/commute costs.
But anything that could allow you to avoid those things…could also allow someone in China to do your job for 20% of what you get paid.
If only people who can afford those things can compete for a job, at least it limits the competition.
Lastly, the best competitive information is scarce. It’s done face to face with no electronic record.
Sorry, but that’s not right either. People living in New York, San Francisco and Boston work remotely, probably more than in other areas. I’m one of them.
If they are working remotely, why the hell are they living in NY, SF, and Boston?
Or do you mean “I can work from home a couple days a week, but still need to be in the office” which is fairly common but still leaves you with needing to be physically near expensive cities.
One spouse needs to be in-person (e.g. doctor).
Because if you bought and own a house in good market, your assets are high. And urban living is great!
My dad has lived in the same house since 1983 and that alone almost makes him a millionaire.
Because most companies that let you telework require you to be in the geographic area and even prove it. Telework in reality is simply a way to shift infrastructure cost to the worker at their expense outside the marginal superstar niche industries which people mistakenly think is the teleworking norm. Norm is minimum wage tied to a home desk with tracking software and cameras that flags your boss anytime you don’t move a mouse for a minute or aren’t physically in front bof the camera and where you have to click “break” for your scheduled bathroom break even at home. But folk like Arnold don’t live in the real world.
I’ve seen advertisements for such tracking software, and I know there are a lot of low end customer service type work at home jobs. So I can believe it.
Very few business decisions are helped by seeing someone on TV. Text and graphics work fine.
This is mostly wrong. “Business decisions” are mostly made, daily, in meetings — where the decision is a) continue as previously planned, or b) so something different.
Most of the meeting is status reports about progress, mostly in accordance with a prior annual plan. Text and graphics are OK, but adding interactive voice, with the possibility of drilling deeper by the decision maker (continue or change), is why decision makers want interactive meetings.
And they prefer them to be face to face — but for the vast majority of continue or change decisions, a presentation is sufficient.
Zoom is excellent for sharing your own desktop, usually power point or excel, plus having small video blocks of others on the call, plus good quality audio, PLUS free for the first 30 minutes!
Plus easy installation, works out of the box. Was often preferred by IBM users over the IBM produced tele-conference stuff.
Zoom is especially great with two monitors, so you can be “working” on one monitor while attending the web-conference / status reporting / powerpoint display. And if somebody else wants to present some point, switching from one presenter to another is simple.
I was at IBM in Slovakia, often working in the afternoons with US IBMers (6 hrs later) and IBMers in India (4.5 hrs earlier) and even China (6 hrs earlier). Time zones are one of the tough issues for remote workers.
b) do something different.
Usually there are face to face working teams coming up with a different alternative plan, and different projections based on the plan, for the decision makers to choose.
A market advantage over gov’t is precisely how often “changes” to the current plan are decided upon.
This was a fantastic interview, infinitely more interesting than the Reid Hoffman interview of Patrick Collison, which for me was infinitely better than any other podcast/vodcast that comes to mind. I’m surprised Kling started with telepresence as his first (only?) topic from this interview; it wouldn’t make my top 10 of potential topics (of more than 100 I’m sure) this conversation touched upon.
First, considering all the talk about the negative externalities of social media, polarization, and BigTech, it is somewhat mind boggling that the active CEO of a $200B BigTech company under anti-trust scrutiny and the active CEO of a a $35B tech startup are having an unscripted conversation about progress and society. The only thing weirder, in my opinion, is having Tyler Cowen thrown into the mix. It makes you wonder if the web/mobile/cloud revolution has not radically changed the meaning of status and/or influence.
The comments here are fascinated but I find myself disagreeing (once again) with Handle, asdf, and P Burgos (HAPB?). I agree (once again) with Mathew Young’s comment. My disagreement with HAPB comes from personal anecdote from working on a software project that spanned the globe. The problem with Handle’s logic is that he treats the remote/local question is mutually exclusive. The project I worked on mixed and matched remote/local tactics. There were four core teams locally located at four core engineering labs (Germany, South Korea, California, Ontario) that had daily face-to-face “scrums” with remote engineers dialing into one of the two english/north-american scrums based on personal preference (morning vs. evening person). This system worked freakishly well, in my opinion/experience. I think this type of system is what Patrick Collison is hoping to achieve with his Remote initiative at Stripe. The project had everything to do with enabling the best technical teams/people in this organization globally working on the most strategic intellectual property in this company’s arsenal.
Handle’s speculation about offshoring has little bearing on the structure and/or process choices made in my anecdote. asdf’s comment about not needing to move to New York is partially right as some remote super stars are extremely valuable but not independent of the core engineering labs which I think Collison was talking about in his Reid Hoffman podcast (he was talking about core alternatives to Silicon Valley and a “Remote” organization). P Burgos’ comment about remote coaches seems off base to me. I can imagine certain specialists working remotely with multiple athletes/teams that compliment local coaches. These remote coaches would use remote data/telemetry to steer training. Think of a modern cycling coach/doctor that is an expert on blood analysis and other power/cardio metrics. The same holds true for certain specialists working with swimmers, runners, rowers, high jumpers, etc. Its not just pros either: anyone with a decent technical knowledge about athletic training can work with multiple high-school teams remotely and simultaneously. Once again, local and remote are not mutually exclusive.
The most interesting part of the conversation between Zuckerberg and Collison, for me, was their insight/experience about engineering teams and tooling. I haven’t done so, but I suspect the comments on Hacker News about these two Collison podcasts and “remote work” would be more enlightening than my own thoughts/experience.
For some relationships you want to build and nurture and maintain trust. For some of these, you want to look the person in the eye, even shake their hand. All this does not have to be in a conference room or in an office. It includes lunch, coffee, ballgames and nature walks and more. Everyone looks for the _blend_ of interaction modes that works best.
If you want to know what is going to be in a press release before its in a press release, and you don’t want anyone else to know or know that you know, nothing beats face to face.
All that x10 if you want to influence what is in the press release.
That can only be done in person.
I think maybe more important than teleconferencing as an alternative to in person meetings is screen share as an alternative to one-on-one in person collaboration/help/training. I use Microsoft teams and frequently use the screen share. This might be one area where the tele-tech is better than in person. I’d rather show someone something through screen share than by going to their desk.
Arnold’s shift of focus on blue collar could be the killer app…
High value educated occupation jobs are just too complex — they require constant coalition building, trust, selling and self-promotion. Most are so competitive that if working remotely makes you 1% less effective you’ll be squeezed out by those who want it more.
Maybe we just hook up people to the internet and let them lie in bed all the day? No need even to wake them up. Everybody, all the time. Wouldn’t that be swell? — bet somebody could make a movie out of what it’s like in that paradise!
Practical example of combining autonomous and remote work from the Australian outback, previously serviced by a lot of FIFO workers (Fly-in, fly-out) in the construction phase. It was very common for the FIFO workers to air commute from Perth, Queensland tropical coast e.g. Gladstone Rockhampton Townsville Cairns, or Bali Indonesia.
I love this post but one concern about the tele-chauffeur: the driver has no skin in the game. I know my uber driver won’t ram the car into a wall. That driver is in my car. How do you disincentivize a remote driver from being reckless? They may be drinking and texting as my car tumbles off the Tappen Zee Bridge.
So just want to understand. Are we proposing that something like better webcams/3D graphics will make telework easier? I would think harder. I prefer a screen share and a voice over phone that I can’t see.
Paul Graham’s latest essay Novelty and Heresy says:
The Zuckerberg/Collison interview touched upon innovating through a focus on teams, tools, and processes. A relatively low impact technical innovation can sometimes be recombined in novel new ways with existing teams/tools/processes that achieve high impact advancements. I think Kling’s/Collison’s example of Zoom teleconferencing is that after years of mediocre teleconferencing solutions, a new solution emerged that was a simple recombination of existing technologies that together created a breakthrough product. In the Reid Hoffman interview, Collison brings up Zoom and Slack as companies that, like Stripe, created breakthrough products/services in what naysayers thought were mature markets.
This is the point of Paul Graham’s quote. Breakthrough innovation is not only about new ideas, it is sometimes about new ideas that are non-obvious or violate a consensus opinion/assumption. Venture Capitalists say that ideas/products/companies labeled as a “joke” or a “toy” are good signs; these are signs that the non-consensus and/or non-obvious rule applies.
Kling’s “blue collar” work seems to involve “machine operators”. We don’t think about it this way but the multiple comments here about screen sharing preference boils down to people who are “app operators”, software applications run on general purpose computing machines. Antipodean Lurker’s Link points to a story about remote operators of giant surface mining machines/vehicles and these “operators” are not at the controls, they are managing mostly autonomous vehicles; its a new job not an existing job done remotely.
The Web/Mobile/Cloud revolution has provided a whole new set of tools. ML/DL (machine learning / deep learning) software and silicon are at the heart of most of the advances in the fields of AI that are mimicking human perception using ubiquitous sensors from the mobile device ecosystem. These new tools can be recombined with existing teams/processes to create novel new ways to work.
IBM did a huge amount of telework for a couple of decades. It is now reducing that, claiming that “Agile” small, self-directing teams are better for creativity.
I think a big part of this change at IBM is as an excuse to get rid of older, more expensive managers, by pushing them to retire (some early) or move to some hub city or branch. So, IBM hasn’t yet done Agile
The idea of Agile seems good: give a business task to a small group of technical folk, who are co-located together, and let them create a solution. Funny sad that the 3 talkers discussed “improve management training”, which is indeed important for developing countries, yet they didn’t mention Agile or other possible “well educated” management team innovations.
Here’s a good recent summary of Agile results (not what it is, that’s another longer ed talk):
https://www.forbes.com/sites/tracybrower/2019/10/06/is-agile-really-worth-it-evidence-says-yes-if-you-do-these-4-things/#1e05a0405488
Doing it right includes these three things, and IBM isn’t yet doing it right:
– ?Embed customers and their feedback in order to continuously improve
– Deconstruct work into small segments and organize effort into short chunks (typically called sprints) in order to get quick feedback and make nimble (agile!) course corrections
– ?? Dedicate people to teams and focus on one project at a time
Also not mentioned in the conference or notes is that ways of doing things usually involve tradeoffs. Face-to-face is usually fastest, but text messages have their own history which is often valuable for later reference, plus less transportation cost, plus a style. Many young folk seem to prefer multiple text messages, like Slack easily allows, over no-history face-to-face / phone-verbal problem discussions.
Engineers on small teams like Slack also for the async ability of having one person have a problem or issue, and getting one or more replies with help — and developing an FAQ (freq asked questions) for the internal team as the team works thru the problems.
Agile, and probably most “higher productivity” innovations, seems likely to be most effective with high IQ folk. And telework in general.
On medicine, there is already some efforts of doctors tele-diagnosing / helping out nurses who are physically remote. In many ways that’s not so different from software call center support to the remote users.
Final Sci-Fi thought on a cute little robot like Pepper, but with a tablet-sized HD monitor to allow visual two way comm with remote tele-operators.
This is happening all over education. I would argue that access to engaging synchronous education may be the biggest impact of a solid video infrastructure and apps like zoom that make it easy
Lambda school has built a better computer science than any physical program because small groups of students work together all day synchronously. The amount of interaction and the intensity of the iteration and feedback is completely unlike what legacy physical computer science programs can accomplish.
Prenda has created micro-schools for k-8 students in peoples homes and all of the academics happen in the cloud. Access to expertise regardless of a student’s location is a huge leap forward.
Outschool is a marketplace of teachers teaching live classes to students on a huge variety of topics. This long-tail learning was previously silo’d based on the local expertise of adults in that child’s area.
Out of curiosity; what is the source of your claim for Lambda’s superiority?