U.S. tech inventors are highly clustered because of idea spillover | Hacker News Reader

U.S. tech inventors are highly clustered because of idea spillover

citylab.com

104 pointsjseliger48 comments

I'm very suspicious of these conclusions based on the dataset. Equating number of patents with "productivity" is questionable and leads to a lot of confounders. For example, most patents are filed by large companies like IBM, which filed 2,000 patents on cloud computing in 2018 alone.

Large companies like IBM value patents more than startups and have dedicated teams that help employees make tons of filings. I would bet this is part of why NY-NJ-PA, where IBM is headquartered, does better than you'd expect. And, Boston, which has no very large patent filers, does worse than you'd expect.

The 12% gain in "productivity" from moving maybe more to do with inventors joining companies that encourage them to file more patents than any real increase in their output.

I would agree, but there seem to be two fields:

"Assignee" which would list IBM and aramonk new york

and

"Inventors" which list the employee names and the city where they individually reside.

IBM employees in other parts of the country show up outside new york

https://patents.justia.com/assignee/international-business-m...

I think the other cause for inflation is that software patents can be rather dubious compared to other fields. It's easier to claim something is a patent when it is really quite trivial (example: one-click shopping). This is harder to in other fields.

I dont recall if it was an official policy or more of a guideline, but when I interned at IBM in college if you had less than ~15 patents to your name you'd never be promoted to manager.

It's not an ideal metric, but I can't think of a better one.

There are historical antecedents the Royal Society (London) and the Lunar society (Birmingham) in the UK.

You can also see this at work in the games industry there are tons of small games companies clustered around Nottingham because of games workshop - its called the Lead belt.

I am shocked by how low Boston ranks and how well Austin does (note: I have worked on both but live and prefer it in the Bay Area). I know patents are being used as a proxy so are an imperfect standard but Austin isn't a huge place and Boston really is a powerhouse.

I'm also surprised, though less so, by NY's ranking -- but it's a huge diverse economy, plus has a lot of lawyers :-).

Austin is the 11th largest city in the US, now (quite a bit larger than Boston, for example). That's kinda big.

It's also one of the fastest growing cities in the US.

That's just not true.

Austin Metro is 2.17M people, Boston Metro is 4.88M. Boston is also part of the broader "Bos-Wash" chain of 50 Million people, while Texas has about 29 Million people.

Sources:

1. https://en.wikipedia.org/wiki/Greater_Austin

2. https://en.wikipedia.org/wiki/Greater_Boston

3. https://en.wikipedia.org/wiki/Northeast_megalopolis

4. https://en.wikipedia.org/wiki/Texas

Disclaimer: I'm from Austin. My friends and family really think it's a big deal... but, it's only a medium deal :)

If you pull in that far for Boston, then you're pulling in San Antonio into the same metropolitan area for Austin. And my point still stands.

I'm sorry, that's just not realistic. The census folks know what they're doing, and when they map out the census areas they factor in the commuting patterns and economic overlap between these areas.

Austin and San Antonio are close by, but there's a noticeable gap, and nearly no one lives in Austin and works in San Antonio or vice versa. Austin and San Antonio don't have the same housing prices, salaries, etc. Contrast that to a place like DFW where the boundaries all blur together and people cross-commute all over the region, and consequently they're lumped together in the census area.

To put this one last way, the census area described as Boston is about 60 minute driving radius around the central city. On the best of times you could make it from downtown Austin to downtown San Antonio in ~80 minutes, and in practice they each have distinct commuting sheds of 45-60 minutes that partially overlap around San Marcos.

Austin is great, and also it's half the population of Boston :)

The bigger question is why is Houston still a gigantic black hole techwise?

Because it's awash in oil money and big companies who want interchangeable cogs for their .NET machines.

It's also about institutional knowledge. It's (relatively) easy to find people in Houston who understand the value in spending $10mm to drill an oil well which has an 80%+ chance to make a single digit multiplier of your money. There's less cultural knowledge of spending that same amount of money for much smaller chance of a much, much bigger payout.

It depends on what you're asking. Apple, Google etc. know that their employees probably prefer Austin so that's why they're there. A lot of software firms servicing Energy companies are based in Austin since its so close anyways e.g. https://mineralsoft.com/.

Related: if you want to see how different urban areas stack up, this is the best list to use: https://en.wikipedia.org/wiki/List_of_United_States_urban_ar...

By that list Boston is #10 and Austin is #37 in the US, which matches how different the two cities "feel" in my experience as well.

Those numbers are from 2010. Austin has grown a tremendous amount since 2010.

I see these false claims about size a lot, especially spread out huge Texas cities. A lot of people claim Dallas is bigger than DC but DC has way more people, especially if you add in Baltimore and or the millions driving down 95 every day.

I was recently in San Antonio and Austin. Both seem rather small compared to DC or Boston.

DC and Baltimore should effectively be added together into one hub and that puts it well into 3rd place and ahead of Dallas for actual urban areas.

Austin is 30th largest. 11th right now is actually Phoenix AZ.

You'd also think that LA would be up there.

I've been thinking that for the 13 years I've lived in LA, and somehow the city still completely sucks at tech. Silicon Beach is basically a joke in comparison to any actual tech hub - the salaries suck, the companies are largely exactly what you'd expect from LA cliches (marketing, fashion, sports, entertainment, etc.), and the success stories are few and far between. These days it seems like there's more going on in Irvine than there is in LA. (If you don't believe me, just look at HN's "Who's Hiring" posts.)

Part of the problem is probably that LA is so close to the Bay Area - it's easy for tech talent to move up the coast or even semi-commute there instead of building a new hub.

I work at the Orange County office of a fairly sizable Bay Area company, and to be honest, I've not had an incredible amount of desire to move to the bay because the OC salary is basically the same (maybe better) after factoring in the difference in cost of living. That and in OC I can live 5 miles / 10 minutes from the office whereas I don't know how far/long I'd have to commute to live in the Bay.

Admittedly, I don't know what the OC startup scene is like, I've got student loans that can't be paid with equity...

As a native and current Angelino, if the city wants to grow tech, it needs capital. We don't have the businesses, can't keep talent, and cost of living is too high to survive without a good salary. Irvine, despite being an excruciatingly boring cookie-cutter place, is slightly cheaper to live and has the Silicon Valley overflow offices.

A lot of talent starts in LA, but it gets immediately poached by literally every other city. If you don't have a Hollywood tech job, it's grim. I really want to stay, but can see a realistic/near future where I leave. (But I'm not going back to Irvine ever again)

Right there with you - it seems like capital could flow the 300 miles south without much problem, but for whatever reason it just doesn't (I suspect it's at least partially related to the types of startups and founders in LA). Irvine is ok if you're a family-oriented corporate person, but for young tech talent it's the most soul-sucking Stepford Wives place this side of Plano, TX.

In my case, I ended up just not working in LA itself for half the time I've lived here. I've spent at least as much time working in the Bay Area, Austin, Seattle, Irvine, and now Portland as I have working in LA because the companies and compensation are just so much better elsewhere.

Their city region labels are odd. The paper is behind a paywall, but it seems like San Diego & Ventura Co should be added to LA, and Tuscon should be added to Phoenix/Mesa. Maybe they're there already, but it feels like not from their choice of labels (not sure what high tech is in Riverside)

I believe they are using census areas, which are constructed by looking at population and then how they connect (e.g. do a reasonably % of people commute within this region, etc).

Ventura County, maybe, but Sam Diego is hours away from most of LA.

From Long beach you can get to San Diego as fast as San Fernando and Riverside.

I'm also surprised by Austin's rank.

I'm from the area, but moved to SF a while back for much better engineering work. I wouldn't mind moving back some day, but haven't seen many jobs or much upward mobility that would make me want to go back. There are a lot of tech companies but they are largely smaller / regional, or are the cost centers of bigger companies (support, back office stuff).

I get the impression that's changing, though. It would be great to see ATX keep booming and reach the level of say, the Seattle tech scene.

Currently, Google has a 35 story building under construction downtown, Apple is expanding its Austin presence both with devs and manufacturing (new Mac Pro to be built here), Amazon and Facebook also expanding hiring here. A look at LinkedIn job openings shows VMWare, Walmart, Microsoft, Dell, Visa, Cloudflare, Mozilla, Auth0, Blizzard, and AMD all hiring here.

I think Austin still lacks HQ anchors of bigger tech companies save for Dell. But the developer presence is absolutely growing.

I moved to Austin back in Feburary and I've met more folks from SF than folks from Austin.

Austin has a deep tech network, especially around certain sub areas.

That’s unquestionably true (UT is amazing, there are research labs, AMD, ApPle and more...) but it’s small. Perhaps I should have been more clear.

Keep in mind NY is massively larger than any of these other places. Per capita, it's actually pretty low.

I'll say it until I'm blue in the face, Boston built the Valley.

Looks like some people haven't heard of Route 128 and not the Johnathon Richmond song either

Going by patents is kind of a weird metric, when so many technology patents are such complete garbage. And what is not garbage is almost entirely corporate-driven.

It's not like people are inventing things in their back sheds and filing patents like they were 100 years ago.

To be fair, the idea of patenting a lot of what we call "tech" these days is probably clustered for the same idea: most people wouldn't think of patenting software unless they're at or around companies that glorify patent recipients and incentivizing more. Still "idea spillover". But for real, that's probably as close as they can get to studying this with easily obtainable, easily quantifiable data.

> that's probably as close as they can get to studying this with easily obtainable, easily quantifiable data

The abstract (I didn't try scihub, but the paper isn't available from his berkeley page either) worryingly blurs "inventors", patent creation, and "[US] macroeconomic benefits".

Approaches to patents vary greatly with industry and business plan. Both of which vary with region. And their correspondence to national benefit is even more diverse.

Number of researchers might serve as a useful cross-check. BLS Occupational Employment Statistics has categories like "Medical Scientists".

But what portion of software innovation is done by "Computer and Information Research Scientists"? What portion is patented? One might easily imagine innovators shifting from dispersed general industry, to concentrated regions and companies emphasizing patents, with an associated increase in patent productivity, but net innovation and economic loss.

Number of IP lawyers per million population needs a lot of correction to serve as a proxy for the economic benefits of innovation.

At a time when we're in a big trade war, in no small part from trying to impose our industrial policy, I wonder if software engineering as a profession is fulfilling an obligation to clearly communicate its experience of just how sensitive impacts are to all the knobs and dials of patent law and policy, and how easily it goes severely net negative. When I encounter economists assuming a general connection between patents and societal innovation across very dissimilar industries, I wonder which of us is failing the other.

Will the existing Bay Area juggernaut ever claw north to Sacramento or south to Monterey, then San Luis Obispo, I wonder.

that could happen with the right investments in public transport infrastructure...

so... no.

Intel has a Folsom campus, and some googling tells me it's [Sacramento] big and does R&D.

I'd like to see this analysis for just independent inventors. From my experience, what really drives patent production is being employed by companies that prioritize patent production... and the places where these industries are concentrated will tend to dominate patent production by volume.

I can't believe the census bureau thinks Bogalusa is part of the New Orleans metro area. The last anyone here heard about them was either the state taking over their finances or water system or that scandal where some reasonably ranking cop had KKK robes in his barn.

It sure feels like they are implying causation where we just really see correlation.

Wondering why the Research Triangle isn't on this list.

Reply on news.ycombinator.com