Programmers don’t become more productive when they move to Silicon Valley or Seattle or NYC.
Why do you believe this? I definitely became a better developer when I moved to NYC. How do you know that everyone else didn’t either?
correlation of wages with housing prices and with wealth is stronger.
I think this is just recursive? Of course wages are higher in places with more wealth. Higher wages causes more wealth. And housing prices can follow, just cause of supply+demand (there’s a higher supply of dollars, so people are willing to spend more of them).
——
Anyway, it’s incredibly effective to take very talented people, and put them together on the same team. They build on each other’s talents, and it becomes multiplicative in a lot of cases. I like the story about the IBM black team as an example (though admittedly it may be apocryphal).
Achieving this requires gathering them, which requires paying them well, which means you need somewhere like SV or NYC where that can happen.
I mostly don’t think remote work is nearly as effective. It massively reduces the compounding effects of having smart people work closely together. But this is just based on my personal experience so far.
Programmers don’t become more productive when they move to Silicon Valley or Seattle or NYC.
Why do you believe this? I definitely became a better developer when I moved to NYC. How do you know that everyone else didn’t either?
“programmers” is too large and diverse a group to meaningfully discuss. For this purpose, it’s sufficient to say “_SOME_ programmers become more productive when able to interact in-person with other workers and experts, which are currently concentrated in a few cities”. And that is a pretty easy claim to make—at least xepo and I assert it to be the case for ourselves.
I’ve done a lot of hiring, engineer evaluations, and related attempts at productivity-measurement for a very large software company. I can say with certainty that there was good evidence that the enforced shift to WFH was a step-change loss in productivity, only some of which has come back. I will also say that it’s NOT evenly distributed—some teams and individuals did recover quickly (and even benefitted). The median and mean was quite negative, though. Standard caveats: measurement is based on imprecise proxies, and Goodhart may have made it even more variable: it was a visible excuse for a performance drop, rather than trying to game the metrics to look good.
I think the net value added per programmer-hour of some open-source projects where everyone works remotely is far higher than anything done at an office for a corporation. Do you disagree about that?
I don’t know how I’d evaluate that without specific examples. But in general, if you think price signals are wrong or “more misleading than not” when it comes to measuring endpoints we actually care about, then I suppose it’s coherent to argue that we should ignore price signals.
I wouldn’t say that “in general” but there are some situations where I do think price signals mean little. For example:
prices of expensive modern art
prices of expensive clothes
valuations of some startups
salaries of many CEOs
America today has large income inequalities, but income inequality < wealth inequality < power inequality. One thing the items on the above list have in common is LARPing by the ultra-wealthy.
Can we measure this somehow? Seems like something someone would have already studied. For all the perceived value of open source, does it actually generate a lot of economic value? Probably, seems likely, but until it’s quantified we’re just arguing intuitions.
My own guess is that open source provides about average value, and the real high value adds come from engineers building things you’ve probably never heard of, like some obscure performance improvement or new feature that increases conversion rates half a percent for some large organization and thus produces tens of millions of dollars in revenue for one FTE quarter worth of work. And for this kind of work, maybe it really does help to be in person, because it requires knowing a large amount of context about the business in order to be able to effect the necessary changes in the code, whereas open source depends more on things that are overdetermined, and so just a matter of someone smart working on them, and thus less coordination is needed.
Sure, there are studies on programmer productivity with remote vs office work, but they don’t all agree, and measuring programmer productivity is notoriously difficult. (Do you use lines of code? Commits? What?)
Here’s one from Baidu that seems decent. According to this, “64% of developers said they were more productive working remotely, compared with 55% in 2021”. But bringing up such studies with xepo didn’t seem like it would be productive, considering the strong statements they made based on their personal feelings.
On priors I think you should strongly expect in-person co-working to produce much fatter right-tails. Communication bandwidth is much higher, and that’s the primary bottleneck for generating right-tail outcomes.
The Baidu study shows slightly longer right tails for individual productivity with remote work, and IIRC others have shown longer tails for remote work as well.
Or did you mean right tails for overall project results?
Yes, right tails for things that better represent actual value produced in the world, i.e. projects/products/etc. I’m pretty skeptical of productivity metrics for individual developers like the ones described in that paper, since almost by construction they’re incapable of capturing right-tail outcomes, and also fail to capture things like “developer is actually negative value”. I’m open to the idea that remote work has better median performance characteristics, though expect this to be pretty noisy.
Oh, I actually think those studies are probably accurate for the thing they’re measuring, which is ”short-term individual developer productivity”. But they don’t really account for “long-term productivity” nor “team productivity”, both of which I think benefit a lot from being in the office. You get an uptick in people’s ability to focus, but downtick in people’s ability to communicate, and both education and coordination are dependent on the latter.
As a counterpoint, consider that ~every major tech company is constantly pushing for people to be back in the office. I know the reddit groupthink about this is that managers are just being dumb, but I think it’s more likely that the individual devs don’t see the impact that working-remotely is having on the productivity of the company over time.
Why do you believe this? I definitely became a better developer when I moved to NYC. How do you know that everyone else didn’t either?
I think this is just recursive? Of course wages are higher in places with more wealth. Higher wages causes more wealth. And housing prices can follow, just cause of supply+demand (there’s a higher supply of dollars, so people are willing to spend more of them).
——
Anyway, it’s incredibly effective to take very talented people, and put them together on the same team. They build on each other’s talents, and it becomes multiplicative in a lot of cases. I like the story about the IBM black team as an example (though admittedly it may be apocryphal).
Achieving this requires gathering them, which requires paying them well, which means you need somewhere like SV or NYC where that can happen.
I mostly don’t think remote work is nearly as effective. It massively reduces the compounding effects of having smart people work closely together. But this is just based on my personal experience so far.
“programmers” is too large and diverse a group to meaningfully discuss. For this purpose, it’s sufficient to say “_SOME_ programmers become more productive when able to interact in-person with other workers and experts, which are currently concentrated in a few cities”. And that is a pretty easy claim to make—at least xepo and I assert it to be the case for ourselves.
I’ve done a lot of hiring, engineer evaluations, and related attempts at productivity-measurement for a very large software company. I can say with certainty that there was good evidence that the enforced shift to WFH was a step-change loss in productivity, only some of which has come back. I will also say that it’s NOT evenly distributed—some teams and individuals did recover quickly (and even benefitted). The median and mean was quite negative, though. Standard caveats: measurement is based on imprecise proxies, and Goodhart may have made it even more variable: it was a visible excuse for a performance drop, rather than trying to game the metrics to look good.
I think the net value added per programmer-hour of some open-source projects where everyone works remotely is far higher than anything done at an office for a corporation. Do you disagree about that?
I don’t know how I’d evaluate that without specific examples. But in general, if you think price signals are wrong or “more misleading than not” when it comes to measuring endpoints we actually care about, then I suppose it’s coherent to argue that we should ignore price signals.
I wouldn’t say that “in general” but there are some situations where I do think price signals mean little. For example:
prices of expensive modern art
prices of expensive clothes
valuations of some startups
salaries of many CEOs
America today has large income inequalities, but income inequality < wealth inequality < power inequality. One thing the items on the above list have in common is LARPing by the ultra-wealthy.
Can we measure this somehow? Seems like something someone would have already studied. For all the perceived value of open source, does it actually generate a lot of economic value? Probably, seems likely, but until it’s quantified we’re just arguing intuitions.
My own guess is that open source provides about average value, and the real high value adds come from engineers building things you’ve probably never heard of, like some obscure performance improvement or new feature that increases conversion rates half a percent for some large organization and thus produces tens of millions of dollars in revenue for one FTE quarter worth of work. And for this kind of work, maybe it really does help to be in person, because it requires knowing a large amount of context about the business in order to be able to effect the necessary changes in the code, whereas open source depends more on things that are overdetermined, and so just a matter of someone smart working on them, and thus less coordination is needed.
Sure, there are studies on programmer productivity with remote vs office work, but they don’t all agree, and measuring programmer productivity is notoriously difficult. (Do you use lines of code? Commits? What?)
Here’s one from Baidu that seems decent. According to this, “64% of developers said they were more productive working remotely, compared with 55% in 2021”. But bringing up such studies with xepo didn’t seem like it would be productive, considering the strong statements they made based on their personal feelings.
On priors I think you should strongly expect in-person co-working to produce much fatter right-tails. Communication bandwidth is much higher, and that’s the primary bottleneck for generating right-tail outcomes.
The Baidu study shows slightly longer right tails for individual productivity with remote work, and IIRC others have shown longer tails for remote work as well.
Or did you mean right tails for overall project results?
Yes, right tails for things that better represent actual value produced in the world, i.e. projects/products/etc. I’m pretty skeptical of productivity metrics for individual developers like the ones described in that paper, since almost by construction they’re incapable of capturing right-tail outcomes, and also fail to capture things like “developer is actually negative value”. I’m open to the idea that remote work has better median performance characteristics, though expect this to be pretty noisy.
Oh, I actually think those studies are probably accurate for the thing they’re measuring, which is ”short-term individual developer productivity”. But they don’t really account for “long-term productivity” nor “team productivity”, both of which I think benefit a lot from being in the office. You get an uptick in people’s ability to focus, but downtick in people’s ability to communicate, and both education and coordination are dependent on the latter.
As a counterpoint, consider that ~every major tech company is constantly pushing for people to be back in the office. I know the reddit groupthink about this is that managers are just being dumb, but I think it’s more likely that the individual devs don’t see the impact that working-remotely is having on the productivity of the company over time.