Hacker News Reader

Comment by adrianratnapala | Hacker News Reader

No idea if Mr. Thiel is right in this decade, but long term silicon valley has lost its main advantage: silicon.

The physical business of building computer parts, especially chips, involves a lot of hard science and deep magic and requires a kind of priesthood steeped in that magic. SV's original success was built on this -- it got rich because it is where the silicon priests lived. But now things have evolved so the it is the Chinese who know how to build the computers and (sometimes) the Californians program them.

Except writing apps is not really deep magic, and is something that can be done anywhere. So SVs advantages move to ever more nebulous things like networking and mentorship -- which are also human universals that happen everywhere.

View on HN

Two elements of 'Silicon' remain vibrant in Silicon Valley, even if the Fabs are all gone: First, the Design software and the design itself is still largely in place here. Companies like Cadence, Synopsis, Intel, nVidia, AMD, and many other chip firms still draw on Silicon Valley's work force to design the most complex chips.

Second, the equipment that is capable of fabricating the most complex chips is almost exclusively developed here. Companies like Applied Materials, Lam Research, and KLA-Tencor make the machines that perform mind boggling feats of physics and chemistry[0]. Those companies are currently dominating 3 of the 5 top spots for revenues globally. One of the other 5 is in the Netherlands (ASML), the other in Japan (Tokyo Electron).

So the actual fabs (along with their Superfund legacy) are gone. But the Silicon innovators are still here and thriving.

[0] https://www.youtube.com/watch?v=4G8wXQGEBrA

Sure, design happens in SV, but "design" is something you can do on a laptop (well, as long as you have a beefy box somewhere else on the network to run the layout step.) It's similar to how NYC is the capitol of fashion design: yes, a lot of designers have congregated there, but any individual designer can still do their job anywhere else in the world without much hindrance.

That's not like how SV used to be: before automated layout, design and fabrication were forced to be vertically integrated—you needed a real fab to iterate your design against. So the designers needed to stay where the fabs were, and the fabs would get built where the designers were.

Uncoupling the two means that the fabs can go where it's cheap, and IC design is now just another kind of whiteboard-driven engineering, rather than a kind of monastic ritual performed within a warehouse-sized machine.

It's a shame that this happened, but you can see how it did.

Those chips were built on the fervor of people imagining what we could someday do with them. They could create a connected world, solve and create so many problems on a global scale...it's not super surprising that their successors went on to do those things.

What is surprising, and shameful, is how investors reacted, buying into the mania hook line and sinker. There cannot be a Bell Labs or Fairchild Semi in silicon valley today, because they could not promise 10,000,000% user growth and a massive acquisition in N years.

It seems like the market equivalent of snorting pixie sticks instead of eating a balanced meal, but I'm not the person with billions of dollars to swing around.

> There cannot be a Bell Labs or Fairchild Semi in silicon valley today, because they could not promise 10,000,000% user growth and a massive acquisition in N years.

Sure there can, they just can't have VC as a part of their business plan.

Thankfully, more and more startups are realizing that VC is not their only option, nor usually their best one.

Good point - they will still need someone to front a lot of money, though. Talent and film-deposition facilities ain't cheap.

I guess they could go with a Bajillionaire Patron, or get funding through usury. Sorry, those aren't even close to fair comparisons, but 'bootstrapping' does seem pretty out of reach these days, for people who don't already have 8-10 figures laying around.

Yeah, I agree. It's not like Bell Labs was ever "scrappy," though.

Ha, yeah I guess they did sort of get the 'wealthy benefactor' thing with Ma Bell.

I've always understood the real "bootstrapping" story to be "we build a successful business around a product we don't really care about in order to fun an R&D department that can build the products we do care about."

Like Google, with ads.

> There cannot be a Bell Labs ... [today]

Google's R+D arms are arguably in the same broad category as Bell Labs was when it started out. We just need to wait 50 years to see if we will view them the same way in hindsight.

It's more likely to be Samsung or Murata or something.

How is Fairchild Semiconductor so different from PayPal? Both have insanely influential and successful alumni. I could have picked General Magic (Pierre Omidyar, Andy Rubin, ...) but they are less well known.

How is Bell Labs so different from Google X? How is Bell Labs any different from IBM San Jose Research (SQL, photoresists, Gray+Agrawal, relational databases, reliable databases, ...)

Silicon Valley has not been about the raw production of silicon chips for over 30 years now.

It's been about software and design of hardware for at least that long.

You're also ignoring the networking advantages (personal, career, and business) that proximity provides. They are not trivial. It's why SV has stuck around post dot-com.

It's almost like the VCs in whatchamacallit road tell everybody else to telecommute, but they themselves don't care much about moving around to keep an eye on the troops.

Lazy bas --- err, I mean, damn efficient, well done, old beans.

(Screw the cattle who have to deal with the fallout. Not adding value.)

An argument can be made that when it was first called Silicon Valley it should have been called Metal Oxide Valley. Originally, SV lived off defense, then made disk drives and now Google, Apple, Facebook, Oracle, ... reign supreme. I think SV will continue to shape shift. It is however, getting kinda expensive.

Andy Grove would agree with you [1]. He argued that a lot of valuable innovation occurs during the scaling/manufacturing process. If you offshore that step, all of the proficiency and innovation that occurs 'at the bench' develops - and is capitalized upon - elsewhere.

[1](https://www.bloomberg.com/news/articles/2010-07-01/andy-grov...)

Reply on news.ycombinator.com