> location you're physically sitting in doesn't really matter
Of course it does.
The benefits of proximity to business clusters [1] is well researched [2]. There is no evidence remote work has dampened that tendency; if anything, as evidenced by AI, the effect seems to have increased.
You're linking to pre-covid studies, that mention some types of benefits (for specific reasons like logistics benefits for businesses relying on physical materials/goos, or physical access to people for the purposes of networking), for some kinds of industries, and also mention that these benefits are not seen for some industries.
> Sometimes cluster strategies still do not produce enough of a positive impact to be justified in certain industries.
Let's take a step back and look at the fundamentals of a tech company who's employees are remote - what are the specific benefits of having a San Francisco office?
I'm linking to studies summarising a century of work. There is no evidence Covid changed this.
Exhibit A for Covid having not changed this is the continuing supremacy of Silicon Valley (tech), Shenzhen (manufacturing) and New York (finance) as industrial clusters that others have tried to replicate (everyone, America and Miami, respectively) and failed.
> Let's take a step back and look at the fundamentals of a tech company who's employees are remote - what are the specific benefits of having a San Francisco office?
Proximity to investors. Proximity to customers. Proximity to a skilled employee pool. Proxomity to acquirers. (A lot of deals happen at cocktail parties and ski trips.)
By the way, I'm not arguing anyone needs an office. Just people physically and and proximate to the cluster.
> Exhibit A for Covid having not changed this is the continuing supremacy of Silicon Valley (tech)
You mean like when China built chatgpt 4 in a weekend and open-sourced it for giggles?
> Shenzhen (manufacturing)
..or you mean how US manufacturing 'clusters' almost completely disappeared/replaced by Chinese ones?
> Proximity to investors. Proximity to customers. Proximity to a skilled employee pool. Proxomity to acquirers.
Right, cause you have to physically pet every investor on the head to fundraise, your remote employees must be in san fran cause network latency interferes with their windows remote desktop workflow, and, for some reason, you have to be physically close to your customers when you're selling your web-only llm wrapper saas.
> You mean like when China built chatgpt 4 in a weekend and open-sourced it for giggles?
DeepSeek R1 is an amazing feat. It's not at the same level as other large frontier models from American companies, just close enough to make them sweat.
> ..or you mean how US manufacturing 'clusters' almost completely disappeared/replaced by Chinese ones?
The value of goods manufactured in the US has never been higher. US manufacturing focuses on goods at the top of the value chain: jetliners, cars, semiconductors, medical scanners, and other advanced electronics. These tend to cluster in a few places - for example, Long Beach is a hub of space and avionics manufacturing, Texas has the "Silicon Prarie," and Boeing in Everett is one of the major employers in the region. Manufacturing has disappeared as a share of GDP, but that's not because we make less stuff.
> Right, cause you have to physically pet every investor on the head to fundraise, your remote employees must be in san fran cause network latency interferes with their windows remote desktop workflow, and, for some reason, you have to be physically close to your customers when you're selling your web-only llm wrapper saas.
You sound sarcastic, but YES. You do generally have to physically be present to make deals. It is obviously theoretically possible to run a world-leading software company fully remote. Despite that, most of them are in-person at least a few days a week. If you want to take advantage of SV's easy VC money, you absolutely have to be present.
Companies are not purely about the numbers. A lot of business is imprecise and heavily dependent on things like just plain LIKING the people you're in bed with, and unfortunately there is no substitute for being in the same room as someone to make a decision like that.
> Manufacturing has disappeared as a share of GDP, but that's not because we make less stuff.
We assemble. The actual parts are largely made overseas.
Because location matters, assemblers in China are able to do better work at a much lower price, see every Chinese EV company, or Chinese drone companies.
Heck in China you can buy a competent Chinese EV motorscooter for less than a kids bicycle in America.
> It's not at the same level as other large frontier models from American companies
Cool story
> The value of goods manufactured in the US has never been higher... Manufacturing has disappeared as a share of GDP, but that's not because we make less stuff.
You're seriously going to try to argue how US manufacturing (which was the greatest share of global manufacturing by far) didn't decline, to support a point about how China (Shenzhen) is apparently an unreplicable manufacturing hub due to 'cluster effect'? It's funny that you don't see the contradiction. For the record, US manufacturing as a share of global manufacturing has significantly declined over the years. [0][1]
"
> The United States' share of global manufacturing activity declined from 28% in 2002, following the end of the 2001 U.S. recession, to 16.5% in 2011
> China displaced the United States as the largest manufacturing country in 2010
> Manufacturing output, measured in each country's local currency adjusted for inflation, has been growing more slowly in the United States than in China, South Korea, Germany, and Mexico
"
> You do generally have to physically be present to make deals.
If you need to be physically present for a meeting, you make a trip. There is absolutely zero benefit to living in close proximity to investors the rest of the time.
> It is obviously theoretically possible to run a world-leading software company fully remote.
It is obviously not just theoretically possible, and many companies do so.
This is the opposite of true. It's actively wrong. Location is almost everything; people move to financial and tech centers for a reason -- it matters where you are and who you know.
> where you are and who you know are very different things
Different but related. Getting a purposeful introduction involves a lot more friction than being invited to someone's home for dinner with their colleagues and contacts.
This is so funny to imply that you (living in East Jesus, Texas) have a better or similar opportunity to me (living in SF) in making more relationships and connections to AI related companies, engineers, investors, customers, acquirers, scientists, etc.
I live a lot further from SF than Texas, yet I've been working fully remote for SF tech companies (among others) for 10+ years.
If I need to meet someone in person, I make a trip (~few times a year)
It's true that I can't brownnose/service random tech talking heads in person on a daily basis tho, which is what I assume you mean by 'relationships and connections' lmao
Okay so what you're doing is contradicting the objective advantages/benefits of living near the epicenter of a specific industry with a purely anecdotal example of 10+ years experience in jobs from said epicenter, with the expendable income to travel (domestic/international) for in-person meetings, then defining networking to a disingenuously generalization because it reinforces your opinion.
What if I were to tell you that you can make meaningful relationships and connections w/o "brown nosing/servicing" and its easier to do so in the center of a specific industry?
I'm giving you a specific example of why it's not necessary to be in any particular location to work in tech, or network, collaborate, communicate with other tech people.
Directly contradicting your baseless assertion about how you have to be in SF for those reasons.
Literally a specific, physical example and you're talking about 'defining networking to a disingenuously generalization' ...
Of course it does.
The benefits of proximity to business clusters [1] is well researched [2]. There is no evidence remote work has dampened that tendency; if anything, as evidenced by AI, the effect seems to have increased.
[1] https://en.wikipedia.org/wiki/Business_cluster#Cluster_effec...
[2] https://www.isc.hbs.edu/competitiveness-economic-development...