HN Theater @HNTheaterMonth

Andrew "bunnie" Huang explains an alternative model of innovation in his talk An Alternative to the American way of Innovation[1].

The over-simplified summary is that instead of worrying about the losing the monopoly protection of copyrights and patents - which are expensive to enforce and often bypassed - consider the new types of innovation that become possible when you have access to the all of the knowledge that was previously locked behind IP monopolies.

[1] https://www.youtube.com/watch?v=S39fhrGjr4U

⬐ jimrhods23

I'm unsure why we should be looking at China for innovation. I can't really name anything innovative that came out of China in the last decade.

Many products/technology coming out of China are based on reverse-engineering or outright theft of intellectual property from European or American companies.

⬐ doctorpangloss

It's great to see someone advocating for alternative points of view with respect to IP.

It's a moot point though. Artisanal hardware is an intellectual dead end. Cost and IP and economics and whatever are a bunch of red herrings.

Hardware standardization is pretty much always superior to fragmentation. I've never heard a hardware engineer advocate for fragmentation! Bunnie Huang should be advocating for the artisanal manufacture of software.

Then again, it really begs the question, why don't we use any software that comes out of east Asia? Or let's say you buy into sales and profit and economics and mom-and-pop artisanal crafts and whatever as being in and of themselves valuable. It's not happening in Shenzhen with respect to software.

The answer is not as simple as localization. Bunnie is having a particularly reductive IP conversation, where there's no law and it's just whatever unwritten cultural norms you personally favor rule the land and can be enforced violently. The IP conversation must be bigger than that, because we live in a world where all the major tech companies, not just Chinese ones, are copying each other's features. And yet, nobody's calling that a lawless Shenzhen world. It's not really about copying or original inventions or any of what he's talking about!

So lemme ask you this: would you rather subscribe to an IP policy in the service of an idiosyncratic electronics hardware community? Or would you rather subscribe to whatever IP theory lets you write good software?

I'd choose the later. That theory is "Whatever the license says," and the diversity of licenses and their social impact speaks for themselves. For that to work, you have to respect the rule of law. Which isn't saying much!

More importantly, lawyers are innovative too. That legal work around OSS licenses has had a much bigger, much more positive impact on the quality of software than a fuck-the-IP-law attitude going on in Shenzhen.

⬐ sanxiyn

> Why don't we use any software that comes out of east Asia?

I do, and you should too. For example, I evaluated mobile-focused neural net inference software and ncnn from Tencent came on top: https://github.com/Tencent/ncnn

⬐ gkanai

> why don't we use any software that comes out of east Asia?

Ruby Language is driven by Matz from Japan. I'm sure there are other projects from East Asia as well that I can't name at the moment.

If you're using a Nintendo Switch or a Playstation, those are also developed by Japanese companies in East ASia.

If you want to look at smartphone apps developed in East Asia, there are many that are very influential such as Line in Japan, Kakao in Korea, WeChat and Alipay in China, etc. Granted these smartphone apps use either Java or Objective-C but they are 'made in East Asia."

In fact, for the China smartphone market, because China blocks most non-Chinese Internet services via the Great Firewall (i.e. Google, Twitter, Facebook, etc.) the vast majority of Chinese apps are made in China. If you don't live in China or Japan or Korea, you probably don't use the apps that are made for those markets.

⬐ askvictor

I suspect that by East Asia, the comment meant China (or hadn't really thought about it too hard).

⬐ speedplane

> Hardware standardization is pretty much always superior to fragmentation.

I think this is changing with the death of Moore's law. In the past, it was almost always better to follow conventional hardware standards because chips kept on getting better extremely quickly. In the 18 months or so it would take an engineering team to come up with some custom hardware, the general purpose Intel hardware would improve so much that it made more sense to wait than to make something custom.

However now, Moore's law is pretty much dead. Processor performance is improving, but it's no longer on an exponential. Today, to gain an extra X% of performance, your best bet is often build your own hardware or ASIC customized to your particular use case. We see this everywhere now, with Apple building their own chips, folks making custom chips for AI and blockchain, and ARM's comparatively customizable architecture beating out Intel's relatively closed one.

With the death of Moore's law, coupled with the unceasing desire to suck out more performance from similar hardware, we're now seeing far more fragmentation than we did in the 90s, 00s, or early 10s.

⬐ _emacsomancer_

> why don't we use any software that comes out of east Asia?

I assume by "east Asia", you mean the People's Republic of China, and not, say, Japan or Korea.

And do you mean software that comes out of Chinese companies?

Certainly I use software developed in China, but usually by individual developers rather than a company.

But think about smart phone software - certainly there is plenty that comes out of China, Korea and Japan.

⬐ sgeisler

> But artisanal hardware is an intellectual dead end.

I agree that this model of iterative improvement won't yield the next big technological breakthrough. But the big consumer electronics manufacturers aren't really innovating either. They might make a product a little bit flatter or give it another camera, but I can't think of a real game changer in the last five years.

Once all the technology is there and just improves iteratively the described model seems to be better suited to find as many local optima in the possible product space as possible (and not just a few big ones).

I could see how a lack of patents would lead to more secrecy in R&D labs. If what you are doing is truly novel and complicated (like developing a smaller node type in processor manufacturing) keeping the research secret would keep you as safe from competitors as patents. And if you aren't (like inventing rounded corners on smartphones) then you shouldn't have gotten the patent in the first place imho.

⬐ ip26

They may not be innovating, but advancing the state of the art is a gradual process, and can by itself eventually open up new opportunity. Google's lauded Portrait mode probably flat out wouldn't have worked on the phones of ten years ago due to lacking computational power, and the steady advancement of x86 has ushered in software worlds beyond an i486, including things like entirely new programming languages that would not work well on 100MHz & 4MB of memory.

⬐ cinquemb

>But the big consumer electronics manufacturers aren't really innovating either.

Yet they continue to attract global capital investment flows that crowd out funding for others (like those R&D labs, because they have their own in house labs with no real pressure to get anything novel to market) to explore something that might yield the next breakthrough.

⬐ archgoon

Well; the entire modern smartphone (post iPhone) industry is only a decade old. Expecting another revolution on par with the first in 5 years seems a pretty high bar to cross.

⬐ User23

> why don't we use any software that comes out of east Asia?

I hear Ruby is pretty popular.

⬐ eitally

Many, many enterprises use software developed in East Asia. There just aren't nearly as many software product companies as in the west. Primarily, afaict, it's because in developing economies like CHina & India (or the Philippines, or Vietnam, ...) there's been enough economic uplift from doing things for other people (IT outsourcing) that tackling the top end of the market hasn't been necessary.

⬐ cinquemb

My problem as a westerner living in Asia for a couple of years, is that the type of retail products that I'm interested in (API's for various things from financial exchanges to consumer goods/services) that are available in the US and are actively marketed, aren't really available to retail users in Asia, and it's really hard to get people to care about it even while they are pitching you to use their mobile app/website.

I know it might not be the best thing to say, but i'm glad that the global economy is starting another downturn with major industries that have provided a boon starting to lose the impact it's had because it will incentivize the marginal person to pursue new opportunities because they no longer have enough economic uplift.

⬐ None

None

⬐ mc32

TrendMicro which is Taiwanese/Japanese -well known and used.

⬐ ohazi

> Then again, it really begs the question, why don't we use any software that comes out of east Asia? Or let's say you buy into sales and profit and economics and mom-and-pop artisanal crafts and whatever as being in and of themselves valuable. It's not happening in Shenzhen with respect to software.

Because the approach that Shenzhen uses for hardware is rarely the same approach that they use with software, and their software approach is generally not something to be emulated.

We may get a lot of wacky gizmos out of China, but the vast majority of them run unmodified or craptastically modified Android/Linux/vendor BSP.

The software model that's closest to what Bunnie describes in Shenzhen is the plain old open-source model. To get good quality software, you need to put significant time and resources into these open-source projects, and let people modify or fork them as they see fit.

Really, the goals are just different. The goal of the Shenzhen model is diversity - "there are so many products out there, what should I make that's a little bit different?" The goal of most software is arguably quality and reliability. There's no shortage of software with silly features, and it's usually easy to add a silly feature to an open-source project if you really want something specific. This is often the kind of software modification I associate with the Shenzhen model.

What's difficult is forking a project over an architectural disagreement and then fixing it, because you have to understand a lot more about how the software works before you can do that, and whatever documentation exists is likely to be in English.

⬐ doctorpangloss

> The software model that's closest to what Bunnie describes in Shenzhen is the plain old open-source model

My point is, that model only exists if you respect the rule of law and specifically conventional copyright. So the thing we can point to as a successful (in some conventional sense, as opposed to idiosyncratically "pro-artisanal" Bunnie Huang sense) demo of his theory exists because of, not in spite of Western IP law!

⬐ sgeisler

You can make this argument when it comes to GPL software, since it forces you to open source your changes if you want to ship them. But does it hold for MIT licensed software (or if you are picky about attribution CC0)? Couldn't the existence of MIT or CC0 licensed open source projects disprove that

> that model only exists if you respect the rule of law and specifically conventional copyright

because these licenses are basically blank checks to do with the software whatever you want? They don't really depend on the existence of any IP laws.

⬐ gred

I think the point he's making is that all software licenses (including open source licenses like GPL, MIT, BSD and Apache) are built on copyright law; your rights as the copyright holder are the only thing allowing you to authoritatively stipulate that your software may only be used in such-and-such a manner, under so-and-so conditions.

⬐ sgeisler

Yes, I understand that, but if there weren't copyright laws we wouldn't need a CC0 license for example, because it would be the natural state of affairs: you can do whatever you want.

And afaik MIT, BSD and Apache just add the duty of mentioning the author, which has no practical implications on the usage of the software. At least I wouldn't care if someone actually writes my name somewhere. If someone asked me to relicense software I wrote as CC0 (and it was formerly MIT) I'd happily do so.

That means that the open source movement doesn't depend on IP laws. Some projects might still depend on it, like the linux kernel (because of GPL) but bitcoin for example doesn't (MIT). Both strategies work in practice.

ADDENDUM: What I want to say is: for many projects IP laws don't have any positive effects, they just add friction by requiring the author to add (and think about) a license.

⬐ jancsika

> I think the point he's making is that all software licenses (including open source licenses like GPL, MIT, BSD and Apache) are built on copyright law;

GP's point is that 3-clause BSD-licensed software makes such a modest requirement upon licensees that it's functionally equivalent to software in an environment where copyright law doesn't exist.

I also assume that means that the GGP is implying that the GPL's "golden rule" requirement is a necessary restriction to ensure a healthy FLOSS ecosystem. Or at least that the FLOSS ecosystem wouldn't be as vibrant if 3-clause BSD were the only type of FLOSS license.

⬐ sgeisler

This model is definitely better at serving niche markets.

For example there are phone manufacturers in China that don't follow the silly trend of making phones slimmer, more fragile and with smaller batteries. Another example are mini notebooks like the GPD Pocket and its many clones which you can (could? haven't looked in a while) only get from china.

I'm a happy user of both product categories and don't think any western company could have been convinced to build these. The market is just too small.

Since I find myself in these small markets quite often (I love robust/rugged devices, phones with hardware keyboards, open hardware, ...) that's a super useful development for me. And I think there are more people like me, maybe with different wishes, but in similarly small markets, that would profit from more product diversity and easier customization of products.

⬐ alanbernstein

First I've heard of that Pocket, it kinda reminds me of this https://habr.com/en/post/437912/.

⬐ johnmarcus

this only works in an ecosystem where the 'race to the bottom' is futile, because you are already at the bottom. Hence why it works in China, no one is going to take your idea and produce it cheaper because everyone is already functioning in the most efficient manufacturing space possible.

In the western world, this is unfortunately not the case. Getting something produced is very expensive because our labor, machine, and real estate costs. Outsourcing makes it cheaper, once you have that pipeline in place. So if I innovate here without protection but pay the expensive price to have produced here, there is a significant advantage for person #2 with the China factory pipeline in place to steal my invention and significantly undercut my costs.

I do like his idea and wish our 'prototype' maker economy was more efficient in the US, but China has paid a high environmental toll to be in that place and it is a price we are not willing to pay here in the U.S.

⬐ repsilat

Software works like this. Language package management and open-source libs are a really close analogue to these little factories and markets full of parts. (I also think we do it with practical knowledge when people change jobs, and I think the short tenures in SV are an undersold competitive advantage.)

For things like UI patterns and non-trademarkable design decisions, I think letting good ideas spread quickly is great for the pace of innovation. Good ideas of those kinds more quickly become table-stakes, and the scales tilt further towards "can execute" and "can consistently innovate" over "had a great idea".

⬐ Haga

Maybe a western maker space should have a co-maker space in the east, constantly connected by a always on vidchat wall. Stop stalking the past.

⬐ sanketmehta

30-40% Westerners/world uses electronic devices from east. Not because they are good, but because there is no alternative to Chinese products. They could easily pay 20-30% more for its west alternative IF AT ALL its available. I think apple and other alikes should hit this opportunity and start making replica (as they say improved version) of mi/oppo/hauwei products. Without compromising on security and quality which is easy thing to achieve. And powerfull apple/google ecosystem would always be plus. Now innovation could only come out of mess, by getting hands dirty.

Thats what Steve might have imagined....apple tooth brush, apple car, apple air/water coolers, apple shoes may be. Making phone bigger and edge to edge won't help. Every product, every domain needs to be explored. Current "WRITERS BLOCKS" of new fresh idea could only be crossed by just going with flow and just make better version of every chinese stuff...Who knows what product might strike the market.

⬐ xt00

One way to think about the current state of hardware design is that we are super reliant upon chip vendors. But in this case the parts are poorly documented or require lots of manual work to get the most out of them, or the documentation is only given to some people, and the interplay between these parts and other parts is unclear. So many people who are ignored by the mainstream companies want that same information and once they get it will share with others. Simultaneously the people who get good support from the companies still work very hard to get their idea working. So they are not motivated to give that away. Is there a huge amount of innovation happening? Sometimes, sometimes not.

So I would say having a culture within the HW community that helps support people who want to use chips that are inaccessible to most people because documentation and support is hard for a small company then that is great.

But advocating that somebody who figures out how to make a novel device that uses a complex set of parts that dramatically changes the paradigm and takes years to develop is then immediately copied then we need to protect that. 20 years for a patent? Maybe that’s too long for some areas.. maybe 5 years but then enforce patents more strongly for obvious copy cats?

⬐ nordsieck

> enforce patents more strongly for obvious copy cats?

You're going to have to explain this in more detail; on its face, it doesn't make sense.

⬐ bsder

> But in this case the parts are poorly documented or require lots of manual work to get the most out of them, or the documentation is only given to some people, and the interplay between these parts and other parts is unclear.

Then I would tell you to quit doing business with such companies. There are lots of companies that are quite happy to give you information (I would cite TI, for example).

If you don't like these kinds of policies, then, for example, do your work on a Beaglebone (TI, with almost everything documented) instead of an RPi (Broadcom, good luck with that).

Now, I will say that sometimes documentation is just terrible. But, sadly, even big customers can't always avoid that.

⬐ jl2718

Pretty sure that’s the guy that scaled a small building to show me the MIT apartment he wanted me to rent with him about 15 years ago. Deja-vu moment. Cool guy; couldn’t find a third.

Do most of these ‘factory’ startups make a profit? How do they get investment? It seems like this is only possible in an export-driven economy.

⬐ 1ta

Patents punishes everyone who is not first. Why would anyone innovate when the rewards are 2 lotto away: 1. innovation itself, 2. be 1st among 7B ?

Patents might have been fine in local economy of 25M 100 years ago but they are definitely are not in global economy of 7B today.

⬐ WhuzzupDomal

I certainly agree with what he is saying; but I believe the American way lends itself to developing better premium products.

⬐ mrtweetyhack

If someone with more manufacturing capability likes your idea, they can freely copy, make, and sell. I don't see how that is a win. Perhaps that is why China is still called a Communist nation even with all the millionaires?

⬐ m0zg

None of the products shown _would even exist_ without American innovation.

⬐ Animats

Remember the "maker movement"? "Maker Spaces?" That was supposed to help restart hardware prototyping in the US. It didn't.

The original TechShop in Silicon Valley was kind of like that. People were making rocket components for the X-Prize. But over time, it turned into a factory for Etsy junk. The concept of "hand made" items made on a CNC laser cutter was a bit much, but it kept about eight laser cutters busy at the San Francisco TechShop. Then Etsy dropped the requirement that you had to make it yourself, and allowed outsourcing. The successor to the bankrupt TechShop has most of the same laser cutters, but they're mostly idle or broken.

People made Burning Man stuff, repaired bicycles, and made furniture. OK hobbies, but not going to change the world or even create a manufacturable product.

I don't know of a single maker space in Silicon Valley with decent electronics prototyping tools. Ardunio hacking, yes. Surface mount soldering and pick and place, no. The commercial maker space in San Jose and San Franciso has cast-off CRT oscilloscopes. About half the right stuff is there, half-broken and gathering dust. I'm waiting to see if the new nonprofits starting up can get their act together and actually open.

What's left of "maker spaces" has turned into "STEM" programs that parents push their kids through so they can improve their resume for college.

⬐ someguydave

Obviously, the reason why "maker spaces" lack proper SMT pick n' place, solder mask applicator and other machinery is that it is all quite expensive if it is up to recent specs and maintained. It's also a low margin specialty, so it would be likely that subscribing to a "SMT Makerspace" would be more expensive than just doing business with a proper assembly house.

Even professional electrical engineering shops rarely build their own stuff anymore - specialized assembly houses do the work.

This is the flip side of efficiency achieved through specialization - most pro hardware "STEM" jobs consist of manipulating computer files in various formats and passing the buck to some stranger to do the actual construction. Unfortunately, this leaves a pretty big monetary gap between "circuit stuff as a hobby" and "I want to make a serious product using the best technologies available"

⬐ speedplane

I generally agree with your sentiment, but I can think of at least one counter-example: Tile, which is definitely a new, U.S. based hardware company. I know the founders and used to work with them at a previous company. We had access to tons of embedded equipment and everyone who worked there had a little maker studio at home (one made robotic espresso machines, another did something similar to Google earth but with hourly weather government imagery, another made specialized text-to-speech software for playwrights).

Maker spaces may not have become a thing, but that's okay as long as the makers are still out there. In the U.S. at least, the true maker space is still in your basement or garage.

⬐ Animats

That doesn't scale. Watch some videos of Huaqiangbei.

⬐ speedplane

>> In the U.S. at least, the true maker space is still in your basement or garage.

> That doesn't scale.

It most definitely can scale. If there is a strong national or regional social norm that encourages tinkering in your current garage, basement, or dorm room, it can reach a far broader scale than forcing folks to make things in pre-defined maker spaces that require memberships and monthly fees.

Maker spaces have benefits, like increased connectivity and collaboration as well as efficiently dividing the cost of larger pieces of equipment. However, they also introduce costs, like monthly fees, forced social connections, and engendering group-think. It can be far more beneficial for a group of students/colleagues/friends to meet in their parent's basement or garage than to each sign up for a maker space.