By James
Pethokoukis and Nicholas Bloom

Urban areas have long been home to top universities, startup companies, and the innovations and opportunities they create. But how strong is the link between cities and productivity? And is geography becoming irrelevant in a post-pandemic age of Zoom meetings and virtual work? Nicholas Bloom recently joined Political Economy to answer these questions and more.

Nick is the William Eberle Professor of Economics at Stanford University. This summer he co-authored, along with Tarek Hassan, Aakash Kalyani, Josh Lerner, and Ahmed Tahoun, the working paper “The Diffusion of Disruptive Technologies.” In the spring, he and Jose Maria Berrero and Steven Davis released a working paper titled “Why Working from Home Will Stick.”

What follows is a lightly edited transcript of our
conversation. You can download the episode here, and don’t forget to subscribe to my podcast on iTunes or Stitcher. Tell your friends, leave a review.

Pethokoukis: The listeners love when I read, so let me read from the conclusion if your paper, “The Diffusion of Disruptive Technologies,” which I think will make it clear to people why I’m so interested in it, and why you’ve come on the podcast.

Policymakers in many parts of the world devote enormous energy to foster nascent technologies, ranging from efforts to support academic research to luring start-ups from other cities and nations. Such an infant industry strategy is predicated on the notion that early advantages in innovation and employment will yield lasting benefits for regions, particularly in the form of high-quality employment.

Now, that
notion that technology hubs matter, that it matters where that stuff happens — we
see that a lot in the United States right now. There are a lot of plans to
create new technology hubs in places perceived as left behind so all the high
tech stuff is not on the coast. This notion is very powerful. Is it a correct
notion?

Bloom: Yes, it is in the short run, but less so in the
long run. So I’ll tell you what we did: We looked at 29 technologies that have
come about over the last 20 years —
things like flash memory, driverless cars, electric vehicles, solar power,
fracking, RFID, etc. — and we looked
at the discussions in earnings calls, we looked in job postings, and we also
looked at them in patents. And we basically tried to track out the lifecycle of
these technologies. And what you see is, initially, when these technologies are
effectively born and enter the commercial world, low and high-skilled hiring
tend to be very geographically concentrated. So let’s think of the cellphone. I
think Steve Jobs first held [the iPhone] up or announced it in, what, 2006?
Thereabout.

Both low and high-skilled hiring for cellphones was pretty
much concentrated in Silicon Valley. The engineers were there, but also the
low-skilled jobs (making them, selling them) initially were concentrated there.
What you see is these jobs spread out, particularly the low-skilled jobs. So
within about 20 years, on average for technologies, low-skilled jobs are
completely spread out across the country. So if you think of cellphones, within
honestly 10 or 15 years, jobs in AT&T stores, repair guys — it was pretty
much spread out across the country. High-skilled jobs tend to stick more in the
place the technology was born, and it takes them about 40 years to spread out.
So you can think that even after 20 years, there are still going to be more
cellphone engineers out in Silicon Valley than other parts of the country.

Is it worth
it for a regional government or national government to try to create hubs in
particular areas, particularly if they think the areas haven’t benefited in the
past from lots of growth in high-skilled jobs?

Great question. So I’m going to split it up into “is
it worth it?” and “can they do it?”

Oh, that was
my second question. Assuming it’s worth it, can they actually do it?

Is it worth it? Yes, in the short run, particularly if your
short run is, as a politician is, it’s maybe 5, 10 years out. So if you come up
with a new technology, the large majority of high-skilled jobs and many of the
low-skilled jobs are concentrated in the county, or city, or region it comes
out of. So in the short run, absolutely. You’re going to get big bursts of
jobs, and money, and investment, and firms. We also have been looking at firms,
and there’s evidence of new firm entry. And firms even re-homing, so moving. So
think of a lot of Detroit car companies, US car companies now have opened up
offshoots in Silicon Valley because of the whole driverless car thing out here.
So yes, definitely, in terms of jobs and employment and income, it’s worth it.
We should come back to how you do it, but of course there are some downsides in
congestion and maybe how much you have to pay to get this. But in terms of
employment, you definitely get a clear short-run boost for employment in those
technologies.

So the
technologies you mentioned previously and the at least short to medium-term
boost of jobs — were those the results of very smart government planners
figuring out which technologies and which places, or was it far more organic? Can
this be successfully done as a conscious, directly planned effort?

As a European, I didn’t think I’d ever hear the words “smart”
and “government” together coming out of America. I used to work in the UK
treasury. I’m in favor of smart government, for sure. In terms of these 29
disruptive technologies: 40 percent of them came out in Silicon Valley, and 20
percent of them came out around the Boston, Massachusetts area. So one thing to
note is they’re incredibly concentrated, and in fact even more concentrated
than patenting. So there’s something about these 29 technologies, which you can
think of as basically commercially successful technologies, that are incredibly
concentrated in really two parts of the US. So then there are other places:
Austin, Colorado, Washington, Seattle, other parts of the country. It’s not
just those two. But those two, Boston and Silicon Valley, are 60 percent
combined. So then if we look in the data, what appears to explain the creation and
birth of disruptive technologies in certain areas? Well, universities, number
and amount of research dollars, endowment size (which is honestly a measure of
how rich and wealthy they are), numbers of PhDs, and numbers of people with a
degree in the labor market.

So effectively, if you have an area that has a lot of
research activity going on in the university and very skilled labor markets,
you appear to have a lot of these new technologies being born. Now, the
question is: Can you create that? I mean, probably, yes. California put a lot
of money into the University of California system. That generated, amongst
other things, Berkeley, which, along with Stanford, really birthed Silicon Valley.
So yes, if you really wanted to have a technological renascence in your area — at
least from a reading of our paper — I wouldn’t say the evidence is super strong,
but my reading of it would be putting money into developing strong universities
for training local grads, undergrads, and also a research base seems to
probably be the most obvious way to go about it.

That strikes
me as trying to create a long-term ecology — a “build it and maybe they will
come eventually” philosophy, versus “we have a plan that, over the next three
years, will radically transform our city.” And if you want to do that, maybe you
subsidize a company to build there or something. The policies you’re describing
seem organic: creating the environment for growth and just hoping it happens.

Yes, I’m not sure very top down industrial policy has a
great track record. Paying company X to move to place Y historically has not
worked out that well. There’s the big thing about Foxconn moving out to
Wisconsin. As far as I know, they’re still negotiating that. There are other
angles. I think Pittsburgh has been held up as a successful model for
industrial rebirth. Obviously they have strong universities. They have Carnegie
Mellon and University of Pittsburgh. They’ve also put a lot of money into
revitalizing downtown.

So coming back to the concept: If you want a lot of grads
and post-grads in your city, you want to make the center of town safe but fun,
enjoyable to live in, good schools, etc. So I’m no expert on the exact cookbook.
I can just say in the data, basically a lot of research universities, a lot of
graduates are very predictive of these technologies. And so my sense would be
that’s probably a better long-run plan than trying to pay for a bunch of
companies to come, because the track record on that is not that great.

You mentioned
Seattle, and if I understand the history, that was a fluky thing. The tech
sector there is credited to Microsoft. Paul Allen and Bill Gates, who were down
in Albuquerque, decided to move back home. And if they had not decided to move
from Albuquerque to Seattle, maybe to some other place, they would’ve created a
tech hub some other place. That was a one-off, fluke thing for Seattle. Is that
what you see in a lot of these places? That there’s some unplanned combination
of events? There was a government role in building Silicon Valley, but it
wasn’t intentional. There’s a random element which defies government planning.

Yes, exactly. I mean I’m not sure if you also read Enrico Moretti’s New Geography of Jobs, but it has a great story about that.

Yeah, I
probably stole it from that.

No, it’s a fantastic book. So yes, although as economists
we like to look at the big picture and look across many, many anecdotes. And so
my belief would be that if Bill Gates and Paul Allen had gone back to Seattle,
but it had awful universities, was horrible to live in, terrible climate, etc.,
they may not have stopped there. If they’d stuck, it’d be hard to employ people
or hard for Microsoft and subsequently Amazon to take off. So Seattle does have
a strong university. It is a nice city. It has a good climate. It’s progressive
and appealing to graduates. The same is true of the Bay Area.

I’m sure there are many startups that begin in many cities
across the US. And some of them grow big, and are successful, and attract other
firms. And others probably don’t. And I suspect having universities makes a big
difference. I’m at Stanford University. I live in Silicon Valley, and amongst
my students and fellow faculty members there is incredible interconnectivity
with firms here. All the time people are going in and out of companies. And the
same is true of Berkeley. So I think the ecosystem that’s led, in a large part,
by universities (and also the graduates they kick out and tend to stay locally),
I think is pretty essential for this.

There are a lot of ideas out there to create technology hubs in places where there currently aren’t any. More in the Midwest, in what we want to call “left behind areas.” We had Jonathan Gruber on this podcast a while back to talk about the book Jump-Starting America, with Simon Johnson. Part of his plan was to send a lot of federal funds to build these regional tech centers around the country. The goal was for there to be tech hubs where it would be ripe, maybe they would have a lot of college graduates, but currently that hadn’t happened. Do you think policy makers can do that? I mean, I understand the political aspect of that, but as far as the goal of actually creating high productivity tech cities and regions that aren’t currently doing that — can government do it?

Yes, I heard the podcast, and I like the Gruber and
Johnson book. In many ways, it’s very aligned with what we’re thinking. So yes,
I think it’s a good idea for the federal government to put more money into
technology and research. And you certainly could target a few cities that have
lots of potential — as in cheap housing, a reasonable supply of graduates,
appealing to graduates as well — and put resources into those. And if you run
it through universities and research centers (I think it’s not going to happen
in one year, but I could easily see in five to 10 years), you’d get startups
and spinoff growth, and you get a mini Silicon Valley picking up. So yes,
absolutely. And that’s very consistent with what we find.

I think the one thing worth noting is it’s not clear you
can do it truly everywhere. So to give you an example: Saudi Arabia has poured
huge amounts of money into trying to set up elite universities, and it’s hard.
I mean, Saudi Arabia is not a very appealing place to go if you’re an academic.
It’s very restrictive. It’s not really an environment I’d personally want to go
to. And so within the US, you probably want to think of somewhere that is going
to attract researchers and academics. And there’s a mix of local political
climate, crime levels, things like that. That means in the very worst parts of
the country, it may be very hard to get this kind of thing running, and it
would take a lot of money. And I think Gruber and Johnson said, “Look,
here are cities, maybe Columbus, Ohio, that are already pretty appealing,
already doing pretty well. You can boost these kinds of cities and they can
really take off.”

Is it your
sense — I’m not sure “market failure” is the right phrase, but I’ll use it
anyway — that there’s a huge market failure? That there’s a significant number
of cities which, on paper, because they have lots of smart people and they have
other livability things going for them, should be doing a lot better than they
are, so government needs to step in? Do you think there’s a significant number
of those? Or is it really a few cities? Some of the plans tend to be fairly
grandiose and, I think, believe they’re going to dot the Midwest with all kinds
of hubs.

Great question. I mean to start off, I’ve been involved in policy in this area for many years, and I know John Van Reenen has often said not everyone can be Silicon Valley. We used to meet with the UK government, and every region wanted to be the next Silicon Valley. So Silicon Fen in Cambridge or Silicon Glen up in Scotland, etc.

Or in this
country, Chicago wants to be the Silicon Prairie.

Yeah, so I mean partly there is raising the level across the US, but only so many places can be a tech hub. Within the US, you could easily have one or two more tech hubs, particularly if the federal government or state government pours a lot of cash into it, because you ignite research in an area and effectively grow America’s tech base. It’s harder to see how you could have 20 of them without competing within the US, so it doesn’t make a lot of sense for the federal government to shift too much activity out of Silicon Valley. What probably makes much more sense is to be additional, to say, “Look, we could increase the spending on universities.”

Other things like R&D tax credits actually can work
pretty well, so that’s something I’ve looked at, and they’re quite successful.
So you could think of universities plus local tax credits. For example, I’ve
been talking to Tulsa, Oklahoma, and they have a big policy about Tulsa Remote.
It’s a pretty forward-looking city. There are places like that that are forward
looking, that are nice to live in. You could easily imagine having a burst.
They’re never going to rival Silicon Valley, but certainly on a local level,
they could really ignite a startup sector if the conditions are right.

Is Silicon
Valley over? That’s the click-bait headline. There are a lot of problems with
San Francisco . . . This will move us into our work-from-home part of our
conversation: The Silicon Valley that we’ve seen over the past generation, is
that in good shape? And if it isn’t, is that important?

I don’t think it’s over. I read a range of newspapers, and
the more right-wing ones, because obviously San Francisco is very progressive,
don’t like it and tend to endlessly predict its demise. It’s true that it’s
incredibly expensive, and that’s a big constraint around growth. Living out
here, I’m not sure personally that people living here particularly want a lot
more growth, because it induces density. And so, in some ways, the market
response, which is to put up the price of property both commercial and
residential, is crowding out growth. And so it’s not declining, but it’s
certainly not growing in terms of population. There’s still a huge number of
startups that are spinning out of it, and I think that’s healthy. I mean, we
don’t want all the startups in Silicon Valley. It seems like a very natural
state of things that it’s very successful, it’s the leading place, it will continue
to be so. But there are other hubs.

I mean, the interesting thing about Silicon Valley is just
how it came about, because ultimately a lot of it comes off Stanford and
Berkeley. And the foundation of Stanford is really pretty random. Leland
Stanford, roughly 150 years ago, bought a vast farm out here and then, for
personal tragedy reasons, ended up bequeathing his money to the university that
has done extremely well. But he probably could have done that in other parts of
the country, and maybe could have done that up in Oregon, and maybe that would
have been the Silicon Valley. So it does highlight, I think, reasonable
climate, maybe good local amenities. But there are probably not too many things
that are essential for having a tech hub beyond having successful universities.

And as I suggested, one reason people are raising the question is that we all seem to be more comfortable, and a lot of companies are more comfortable with people working someplace else. That brings me to the second paper, “Why Working From Home Will Stick.” How much will it stick?

Oh yeah, so this is with Jose Barrero and Steve Davis. Working
from home is going to stick, but more at the extensive margin. So just to
explain that: Before the pandemic, roughly 15 percent of Americans ever worked
from home for a full paid day, and that accounted for 5 percent of working
days. So basically it really wasn’t a big deal. During the pandemic, that
exploded to 50 percent of Americans, and probably pretty much most people — me,
you, pretty much everyone listening to this podcast — are probably working from
home or at least have for chunks of the pandemic. The question is: What’s going
to happen post pandemic? The dust is clearing now. I’ve probably talked to 100,
200 firms, and running a lot of surveys, it’s pretty clear post pandemic, the
setup’s going to be hybrid.

So most employees, 80 or 90 percent of us that are
currently working from home, will go back to the office, say, three days a week,
and work at home two days a week. So in terms of the extensive margin, will we
still be working from home post pandemic? Pretty much everyone that is now will
be post pandemic. But rather than doing it five days a week, typically we’ll be
doing it two days a week. And for most people, that’s actually a pretty happy
medium. They’re getting a bit depressed and lonely at home, they want to go in,
they just don’t want to go in all five days.

So I
understand why a lot of workers would like it. But are they productive? I
assume that a lot of companies think they are or they wouldn’t go to it. Or
maybe it’s a tight labor market and that’s why they would accept that. But from
what you found, are people as or more productive?

Yeah, so hybrid’s basically a win-win, and you asked about
productivity and what workers want, so I’ll just start with productivity. So on
productivity: I got into this by running a randomized control trial on working
from home in China 10 years ago now. We randomized volunteers by even and odd
birthdays, and you found people that were randomized into working from home for
four days a week were 13 percent more productive. And so we were amazed. We
were expecting the reverse, in all honesty. The company doing it basically
thought they’d goof off but thought they’d save on office space. Turns out if
you look into the data, about 4 percent of it comes from them being more
productive per minute, and the reason is, it’s quieter at home.

There were amazing anecdotes when you talked to people, like the person that told me the person in their cubicle next to them clipped their toenails at work under the desk. They said it was really disgusting. So the office is noisy; home is quiet. That’s 4 percent. The other 9 percent is folks at home just worked more minutes because they took shorter breaks — honestly shorter toilet breaks, less lunch breaks — they had less delays, where the bus broke down or the car wouldn’t start. So that was that study. There’s a lot of other papers out there. Harrington and Natalia have a paper I’ve just seen. Two Harvard grads show a similar finding. There’s a number of papers. They’re not all universally positive, but I would say that the large reading on the literature, and our survey data, seems to be working from home increases productivity.

The thing to bear in mind is there’s actually two types of
productivity: the short run and long run. So it seems to be particularly
helpful for short-run productivity, basically doing the same thing you’ve done
before. So in terms of repeating tasks, it seems fine, if not better, doing
them at home, because it’s quiet and you have more time, and you save on the
commute. The problem seems to be what I would call long-run productivity: being
creative and innovative. And certainly from anecdotal evidence — and in fact,
we’re following up with an RCT in China on engineers now that are creative in
terms of developing new products — there’s more concern that certainly five
days a week working from home could hamper long-run creativity.

What happens
to people talking to each other, whether it’s at lunch, the water cooler, what
have you? Exchanging ideas: “Oh, did you hear about this?” Are two
days a week enough for all those interactions?

What I’d call the plain vanilla hybrid is three in the
office, two at home. So Apple is ingenious at coming up with simple, easy to
understand products. As it is with their plan: They said Monday, Tuesday,
Thursday, you’re in the office; Wednesday, Friday, you work from home. Fair for
the whole firm. So then is that enough? Well, it looks like you need to be in
the office for basically bigger meetings, maybe informal chats, coffees,
brainstormings, trainings, client events, etc. But that, for most people,
doesn’t take up more than 60 percent of their time.

It seems home is better for reading, writing, preparing
presentations, often one-on-one meetings. People report one-on-one meetings
actually work very well on Zoom. It’s not really much better in person than on
Zoom. Take Apple: Monday, Tuesday, Thursday, you schedule all of your meetings
and events and trainings on those days, and then Wednesday, Friday are quiet
days. Folks go home. They save on the average of an hour commute each day. And
it’s quiet, and therefore they tend to get more done on the quiet stuff on
those two days. So yeah, three days in the office is enough. I would be nervous
about one day in the office or two. You’re right that that’s trickier.

And does it
matter for worker productivity, if they’re brand new workers? I mean, certainly
my concern is that someone starting a job without considerable time in the
office just misses out on a lot of understanding the culture, mentoring, all
kinds of soft skills that people gain by being in person over a number of years
that a new, young employee will have a very hard time picking up.

Yes, definitely right. And so the advice which I’ve been
giving to firms is for new hires, or maybe even people that move positions, for
the first six months or year, you could think of them coming in an extra day. I
used to work at McKinsey and they had this big training thing for the first
year. You can imagine if you joined a firm, you have three days a week in the
office and in your first year you have an extra day, I don’t know, let’s say
Friday. And your whole cohort comes in, and different managers come meet you.
You basically build some connectivity with the firm, have some extra training.
But yes, as time goes on, it tends to be easier. Remember: three days a week in
the office is quite a lot.

Currently, when we go into the office, we’re not spending
all the time in meetings and chatting to people. A lot of it is sat in what I
call Mad Men-style corner offices, whereby you’re on your own working. That
kind of thing is just better done at home. So the idea is you have your three
days in the office — exhaustingly social — but you’ve got all your social and
big meetings out. And then the other two days, you don’t have to shave, you
don’t have to put on fancy clothes, you don’t have to commute. It’s a lot
easier. I mean, I’m talking to you here in a track suit and T-shirt. I probably
wouldn’t be wearing that into the office.

I want to
finish off with just a few questions about productivity growth, one of which
comes from the conclusion of the working from home paper. The productivity
gains from working from home may not show up in the official government
statistics. And let’s focus on both the gains from less commuting, but also the
gains from me being able to have business meetings without doing as much
cross-country traveling.

Yeah. It depends a lot, particularly on what you measure,
in terms of hours. So just think about it. Imagine the average American
commutes one hour a day, and imagine I now work at home. I produce exactly the
same amount, but I stop commuting. That prior commuting time is effectively
work time, and for most people, that’s not particularly fun. But now I’m
producing the same amount through spending five hours less a week commuting. Or
maybe just I’m more efficient at home, but I work a shorter day. So the issue
is partly that national statistics don’t tend to have precise measures of hours
worked. Most people put down their shift hours, whatever it is, and they ignore
commute. I think most of the increase will show up in raised output.

I mean, it’s a really interesting question. I was
discussing this actually this morning at length with some co-authors. In the
US, we currently have GDP above pre-pandemic levels, but we have six million
less jobs. So of course, if we’re producing more with six million less people,
our productivity is substantially up. So the pandemic has seen a surge of
productivity, and that’s against the backdrop of secularly declining
productivity growth. So America’s productivity growth in the decade before the
pandemic was pretty terrible; it was less than 1 percent. So is the pandemic a
new normal? I’d say I’m slightly skeptical. Working from home is great and
maybe kicked up productivity 2 or 3 percent, but I don’t think it’s done much
more than that.

So my guess is some of the burst in productivity we had now will unwind. Some of it is because you can get delivering services frankly at lower quality, like restaurant meals delivered to your home. But once we have to go back in person, all those waiters are going to get reemployed, but you’re going to have the same sales. I don’t know, it’s an amazing phenomenon. The productivity burst is incredible as an economist because if you remember Kydland and Prescott’s famous paper that effectively helped them win the Nobel Prize, it pointed out how recessions used to be periods of negative productivity, really bad periods. Now, we’ve had a recession, the pandemic, with a huge positive productivity burst, so it’s pretty weird. I wish it will stick, but I’m not that optimistic, I have to say. Beyond working from home, it’s hard to think what the pandemic has done to generate such a big increase in productivity.

Well, another story about better productivity growth ahead: You see the phrase “we’re going to have a new roaring ’20s.” I think some of that might be based on hopes for work from home, and maybe the infrastructure plan — somehow that’ll boost productivity. But they’ll look at things like new technologies, whether it’s machine learning, or SpaceX, or CRISPR, and there’s this emerging case that that decade or more of weak productivity growth is at an end. We’re going to have maybe a decade of much higher productivity growth. I noticed that Robert Gordon and Erik Brynjolfsson did a long bet on this very idea. And you co-wrote along with Charles Jones, John Van Reenen, and Michael Webb, a fantastic paper about ideas getting harder to find and how we have to throw more resources at those to get those ideas. Do you think that we’re going to have higher productivity growth, for whatever reason, in the decade ahead? And do you think things like AI will make it easier to find big game-changing ideas?

So I would love for there to be increased productivity
growth. My suspicion is probably not. So one thought is that I don’t think the
pandemic has made it easier to have faster productivity growth because, in many
ways, these game-changing technologies, take AI — you could have done AI pre-pandemic.
There’s nothing about the pandemic that’s made AI particularly easier to use. I
mean, maybe you could argue it’s made it more appealing to have robots rather
than people to, say, serve your coffee, but that incentive is already there. In
reverse, the pandemic, I know from Stanford University, has been hugely
disruptive. It closed down a lot of the labs, still the social distancing
capacity is down, et cetera, et cetera. So it could be we have a productivity
revival. If we do (again, apart from working from home, which is maybe a one-up
lift of 2 or 3 percent), I don’t think the pandemic itself would generate a
permanently higher rate of productivity growth.

But if we saw one, the most likely reason would be that in
2018 or 2019, we had all these technologies that were ripe for expansion — AI,
computer vision, driverless cars — and finally the light switch gets turned on,
or something clicks, and they finally take off. I mean, just to set this in the
course of history, productivity growth was roughly zero basically until the
industrial revolution, and then it suddenly starts growing from roughly zero up
to about its peak of around 2 percent in 1950. And then it’s been declining
back down since 1950 to roughly less than 1 percent pre-pandemic. So it can
change direction. It just hasn’t done that often in history. So I would
probably bet against there being a productivity turnaround, but it’s not
impossible.

Have you
thought about AI as a super research assistant? As an invention to help find
other inventions, to get at that “ideas getting harder to find”? We only have
so many researchers, but maybe AI will amplify or complement those researchers
and we can find more big ideas.

Yeah, exactly. So you can almost call this the Skynet view of productivity growth. In The Terminator, if you remember, Skynet eventually figures it out and kills all the humans, more or less.

Instead of
killing humans, find great ideas. That’s the difference.

I mean this idea has been kicking around that when computers get good enough, computers will come up with new ideas, and then things will explode. Maybe Ray Kurzweil is the person that came up with Singularity, but there’s been a bunch of studies looking at it. Again, it could be. The problem is often when people look at the micro-studies, you often need humans at some point in this chain. So a computer can come up with some great idea, but it doesn’t quite translate or it needs to be put in practice. So there’s plenty of potential. I don’t want to seem too pessimistic. Just looking back at what’s happened since the development of the personal computer, the internet, etc., none of these have really generated dramatic increase in productivity growth.

I find it a bit odd, to be honest. I spent a long time
trying to figure out why they haven’t, but I mean they just haven’t to date. So
yes, of course, we could have a turnaround. But there’s nothing I’ve seen that
says AI is radically different so far, but one thing is true is it’s hard to
predict productivity growth too far out. So I’d be surprised if there’s a
radical turnaround in the next five or 10 years, but it’s like the weather: It’s
very hard to predict the weather more than two weeks out. I wouldn’t really
place many bets on 10 years from now, because AI at that point could honestly
have totally taken over or we could be down to zero growth because — I don’t
think we’ll ever run out of ideas — but something nasty down the turnpike could
slow us down even further.

My guest
today has been Nicholas Bloom. Nick, thanks for coming on the podcast.

Thanks so much for having me. It’s been a great discussion.

James Pethokoukis is the Dewitt Wallace Fellow at the American Enterprise Institute, where he writes and edits the AEIdeas blog and hosts a weekly podcast, “Political Economy with James Pethokoukis.” Nicholas Bloom is the William Eberle Professor of Economics at Stanford University.

The post Should government create more Silicon Valleys? My long-read Q&A with Nicholas Bloom appeared first on American Enterprise Institute – AEI.