Comment by samstave | Hacker News Reader

Comment by samstave | Hacker News Reader

Interesting - the lure of "making something" in Shenzen has been a little seed in the back of my head for years... (one of those secret little pipe-dream-fantasies that pops up every once in a while and consumes 'what-if' cycles, but quashed by reasonable-rational-realist-me, but entertained just the same)

---

That said, one of the adaptations of this little fantasy that I have had, and wonder about how feasible it could be is;

What if there were a group effort with a communal goal of accomplishing X in China...

An informal start-up-market type model where you have agents in some silo categories, but all working together.

Lets assume that the silos are "Design/Eng | Finance | ops | mfr-execution | sales"

So you have local US agents who collab on a design of a thing - you have backers (can overlap), and ops facilitators who focus on where when how it will be sold -- and then you have mfr-execution agents who actually go to shenzen and get the deal done and made... finally a sales org.

Kinda like a communal quirky?

It would be better just to build a Shenzen in the US.

Now THERE’S an idea. What are the major hurdles to overcome in implementing this?

EDIT: Looks like the efficiency comes at a high price.

Relevant (reverse) HN-AMA: https://news.ycombinator.com/item?id=18629299

---

Welp, about 50 years of cheap, environmentally catastrophic manufacturing (lack of) regulations, a vast army of poor/slave-subsistence-level farmers who can CHEAPLY provide food - no IP laws, a mass ability to gather/import raw materials - hazardous working conditions, cancer, cancer, cancer - slaves, and a gumption of a population with a striving for life that is not nearly as lazy and entitled as anyone else on the planet.

Just a start...

OHSA, intellectual property laws...

Convince the people with all the money that building a giant manufacturing center would have a favorable ROI.

No, as much as I dislike Trump, a nice tariff could solve a lot of problems long term. Its after all how this country was built. Put in more economic terms, you should be taxing things you want to discourage. Taxing income/work, while providing free trade with locations that have lower tax structures _WILL_ force jobs and industries offshore.

OTOH, a nice fat tariff on things you want to manufacture, and suddenly manufacturing them at home starts to make strong economic sense. Of course as evidence suggests that you also need a large enough market to assure that your not just paying 100x as much to support some inefficient factory.

Or you just decide you don't need to manufacture that thing after all. There are a lot of situations where "redesign so it doesn't need that version of the part" is cheaper than "build a US factory, train workers."

If tariffs worked on their own, Brazil would be an electronics powerhouse. https://thenextweb.com/la/2012/09/30/from-brazil-cost-brazil...

The NAFTA market is literally 10x the size of Brazil. Its not the same thing.

Worse, tariffs tend to work if your protecting an existing industry from dumping/etc, not so much if your trying to grow one.

You can’t really compare the US/EU/China to individual smaller countries. There’s a scale above which the economics are quite different.

labor laws, wages, safety, regulation, cost of living

Pollution. Regulation. Investors. Unions.

https://www.popsci.com/america-before-epa-photos

”just"

It's not like we have to re-invent the industrial revolution, we just need to spend the money to set up shop.

It's just that the West has decided that allowing cities to grow is a bad thing, and the exact opposite of that is what powers economic growth, in Shenzen and certainly in our own history as well.

In what way has “the west” decided it is a bad thing to let cities grow?

Perhaps it could be some kind of shack where radio components could be found...

Really though I think fast shipping will suffice for most people. Once we get delivery down to under an hour, in city, with reasonable component prices, we should be able to source what we need easily enough. I'm imagining something like jimmyjohns meets sparkfun (and now I'm drooling).

I have a theory about what killed the radio-shack star:

PACKAGING.

Fucking packaging kills "hobby" interests.

Radio shack became "random aisle from fry's electronics" pretty easily. Fry's is complicit in the murder of radio shack -- but the fact that they individually began packaging everything and selling things as expensive individual components in blister/cardboard crap - rather than in a much more organic manner really limited the Hacker energy.

They attempted to bolt themselves to the "charging cable industry" -- (which deserves its own place in tech hell) -- and that failed.

They were then reduced to a charicature of themselves when they were selling 'MAKE::' kits...

I lament for the sould of what Radio Shack once could have been...

Radio Shack had an identity crisis for much of its history and going back to at least the 70s they were as much about selling mostly crappy audio, toys, etc. as they were about selling components. I suspect big box electronic retailers played as big a part in killing them as anything else. But there was also a period, during which Heathkits went away, when it became much more difficult to do interesting stuff as a casual electronics hobbyist at home.

Things like Raspberry Pis, Arduinos, 3D printers, Linux, etc. arguably made a new generation of hardware hacking possible but those came later.

Are you trying to say that the TRS-80 wasn't a top-of-the-line luxury computer?

The TRS-80 at its peak was actually a pretty logical outgrowth of Radio Shack's ham radio and electronic hobbyist business. It outsold the Apple II at one point, but then the entire market basically shifted to Apple and the IBM PC and clones.

I think the core issue is 50’s tech simply had vastly fewer parts. Now days even covering every major type of LED takes significant shelf space.

The way around this IMO was if they had added a in store pickup + some large regional warehouses. Free/Discounted next day shipping to the store of your choice or more to your house, could have really enabled hobbyists without costing the company much. In larger markets they could have even done some same day shipping to stores.

I think this is a little nutty. This business has been subsumed by digikey, findchips.com, mouser, etc. Do you really have the time or inclination to get in a car and rummage through bins? Isn't designing, simulating, or even reading data sheets a better use of your time?

For me, trying to put myself in the mind of an aspiring/learning electronics hacker -- yes. I want to rummage through bins. I want to touch the components. I don't want to stare at a computer screen trying to conjure up the proper search terms to find what I'm looking for.

Later, as I get more advanced, yes, I can see the online vendors making more sense.

The thing you're touching on here I think is also some degree of freedom for iterative and spontaneous revision/change. A searchable catalog is great when you already know what you need. A catalog is also great when you have no idea what you need, but there's a huge delay in the feedback loop with a catalog compared to bins. To empower a discovery process during creation, it helps to have "saturated space" or something akin to untold quantities and varieties of "bins".

I too love rummaging through my bins and wondering what is possible. I think the hacker mind just works like "I wonder if X could be made to do Y..."

Wisconsin is trying to do that and it's not going well.

https://www.theverge.com/2018/10/29/18027032/foxconn-wiscons...

That was a political boondoggle from the start. What Wisconsin was trying to do was to become as corrupt as Illinois, and that's going quite well indeed.

Kenosha-Racine just did not have the number and kind of workers required for the original promises to hold water, and the subsequent revisions only got more ridiculous. For brainy, high-tech work, the only nearby choices are Madison and Chicago. It has always been hyperbole and lies, told to sway votes. The powers behind it didn't expect to lose their statewide offices this quickly, though, so I'm not sure if their plan for the Foxconn site is still viable.

"It would be better just to build a Shenzen in the US"

Probably not.

Imagine if you went to a flea market, but instead of comic books, hot dogs and used sofas ... you found all sorts of cheap, exotic component parts. Anything. Everything. For 10's of cents each.

I'm close with a team making a product, and there are a few 'custom-ish' things they need, like a tiny microcontroller with camera that does some very basic processing as a sub-module. In America - that's a big deal. Hire a team, get some researchers to consult, do some iterations yada yada.

In China some guy in a ratty, dirty, poor little corner shop has it! For like 40 cents in units of ONE!

There's a massive 'secondary industry' of people making 'not rocket science but not trivial' gadgetry very fast, and on the cheap that is not going to be duplicated anywhere.

Add in the tons and tons of cheap labour, with few labour laws.

Add in the 10 hours 7 day crazy work weeks.

No - Shenzen will not be duplicated.

That said - it may be possible to expand the Valley in a more 'automated' 'high end' way.

For example: 3D printers that are more amenable to 'short production runs' instead of just prototypes. Low cost / fast turnaround injection moulding. Quick turn around on smaller components.

This might make it better for American firms doing quality high tech to churn out faster prototypes, first production runs, and later automated large runs.

Doing this would take some champion leaders, maybe Cisco, or Google or a few others setting up shop, proving the way for others to follow.

For example: 3D printers that are more amenable to 'short production runs' instead of just prototypes. Low cost / fast turnaround injection moulding. Quick turn around on smaller components.

This already exists in the North American market, I.E. Protolabs

Funny you should mention them I was literally speaking to them this week :)

Protolabs is nothing special really.

The tech everyone uses for 3D printing, for quick turnaround injection moulding ... it's mostly the same old stuff.

3D printing is only good for a few units, unless what your making is very small and very expensive.

Custom moulding is still a big leap - and if there are any iterations (and there will be) it's not cheap.

And all of the 'fine' stuff is still done by hand.

So I see 3 roadblocks:

1) Rapid iteration and availability of component parts 2) True cheap way to get from 10 to 20000 units i.e. before true mass production 3) 'programmable robot hands' i.e. a very easy to program robot arm/hands that is nimble enough to fold and cut fabric, assemble arbitrary things, wipe/clean surfaces.

Well, you can make 100+ high-quality plastic parts a week on a $500 SLA printer (if you can fit at least 5 on the bed, vertically), and the price per part for the material could be below $1/part (the goo costs $70 per liter). At least in theory, that's my plan in the next weeks. PCBs are no issue in smaller quantities, although the setup cost for PCB assembly will add to per unit cost if you are doing less than ~500.

Here's an example, although perhaps a little big: https://www.youtube.com/watch?v=RDUn3CoVmuc

Looks quite saleable to me. Maybe sand blasting would improve the finish.

The real plus China having Taobao.. like eBay, but with every imaginable component you could need, and five different Chinese clones of it at 20% of the original component cost, all available at wholesale prices, in quantities of 1 or 100,000. The Shenzhen markets are still active, but shrinking.. you can usually find more variety, faster, cheaper, online. You can hand your BOM over to an agent on Taobao, and he'll get the parts together for you and ship them over.. it might cost 1/4 of the price of Digikey.

Yes, if the parts are small and don't need much finishing ... that just might fly. Good luck with that I hope it works for you.

Off the shelf electronics are good for some things ... not so much for others.

Glad to see someone trying it!

Yes, but in my experience, ProtoLabs is the exact opposite of everything available for <$1 -- they specialize in rapid turnaround, and they want all the money for that kind of rush.

I also found that they aren't all that amenable to finding the most efficient route using stock material. E.g., on one job, I just wanted a stamped or laser-cut hole pattern in stock aluminum, so I specified the stock thickness in that field and commented in the notes what I was seeking. I got back a quote for basically CNC milling the exact part out of billet. Lovely, but completely unnecessary, and it didn't improve chatting with the rep. I would up fabricating the parts out of carbon fiber and CNC cutting the shapes/holes myself (all in-house).

So, yes, I'd say that the Shenzen flea-market mentioned above could be really cool, is definitely not here yet, and is sadly unlikely...

(Edit: typo, dropped words)

> Imagine if you went to a flea market, but instead of comic books, hot dogs and used sofas ... you found all sorts of cheap, exotic component parts. Anything. Everything. For 10's of cents each.

Sounds like how Silicon Valley used to be...

Yes, when Weird Stuff Warehouse, HSC, and Haltek were still in business. Steve Jobs bought his first oscilloscope at HSC.

You'd need to have a way to replace humans with machines and do so both flexibly and at-scale. The big competitive advantage of Shenzhen is people - and not just unskilled labor, but a lot of skilled tradespeople & engineers that can design & build electronic components to spec very quickly. The U.S. just doesn't have the density and skill level for that - the closest you might get is Silicon Valley for software, but even here the population density is < 1/4 of Shenzhen, a relatively small percentage are skilled engineers, and we've somewhat overspecialized on Internet, mobile, & big-data software, making it difficult to innovate in other fields without having an extremely rare combination of skills in one individual.

In theory things like 3D printers, pick & place machines, and better CNC millers all controlled by software could help with that, but interest in hardware hacking in the Bay Area has gone down and a lot of the professional electronics manufacturing moved to China.

"but a lot of skilled tradespeople & engineers that can design & build electronic components to spec very quickly"

Yes this, is what I failed to mention.

You'd need to have a way to replace humans with machines and do so both flexibly and at-scale

Isn't this something that Elon Musk is working on? I think I read somewhere that one of the reasons his car factory was so expensive to build is because the robots have the capability to do more than just one thing.

Hmm, interesting thought: what if the reason Elon Musk is overinvesting in robotics with Tesla is because he believes that the initial Mars colonists will be extremely labor-constrained, and will need robot helpers to survive? Current plans for the BFR and Mars Transporter indicate an initial crew of 12 people, most of whom will likely be highly-trained specialists. If you're building a base around 12 people who are likely biologists, doctors, mechanics, engineers, and rocket scientists, who's gonna clean the toilets and grow the food? Plus robots have the big advantage of not needing food and being able to work outside in the thin, freezing Martian carbon-dioxide atmosphere.

https://twitter.com/elonmusk/status/984882630947753984

> Add in the tons and tons of cheap labour, with few labour laws.

> Add in the 10 hours 7 day crazy work weeks.

These advantages seem like they'll expire as China becomes more competitive economically.

> Imagine if you went to a flea market, but instead of comic books, hot dogs and used sofas ... you found all sorts of cheap, exotic component parts. Anything. Everything. For 10's of cents each.

How many types of sensors / components etc. are even conceivable, no matter how exotic? It seems like someone could replicate that in the US with a warehouse and sourcing. These things are extremely cheap in general; it's not like trying to replicate the VCs of Silicon Valley elsewhere.

> For example: 3D printers that are more amenable to 'short production runs' instead of just prototypes. Low cost / fast turnaround injection moulding. Quick turn around on smaller components. ... Doing this would take some champion leaders, maybe Cisco, or Google

This seems like a great idea; pair some small fabs with a big ol' warehouse stocked with every random component / sensor under the sun, and it seems like it would work out totally fine.

"It seems like someone could replicate that in the US with a warehouse and sourcing. These things are extremely cheap in general;"

There are an infinite number of module combinations. They all vary in cost, quality, precision, speed, size, materials, performance, suitability for manufacture, and subtly in whatever it is they do.

A commenter above mentioned the vast number of professionals with a ton of know-how in this area, making the 'time to tweak' or 'from nothing to module' and also 'optimized for manufacture' very rapid. There's nothing remotely like it anywhere.

Not really, we need to figure out how to stop making stuff and contract economically for our grand kids to have a chance

I'm surprised nobody countered you on patent law immediately. Shenzhen's biggest advantage for a R&D guy like me is that nobody cares about patents. Just come up with ideas, copy ideas, remix, whatever... it doesn't matter because whoever has the best product for their market will make the money. It can also serve markets better when vendors with patents refuse to give customers what they want. The iPhone clones letting you swap your battery easily are a good example.

Nice documentary on Shenzhen if you want to learn more about its history, culture, and advantages:

https://www.youtube.com/watch?v=SGJ5cZnoodY

In the U.S., you get this instead:

https://en.wikipedia.org/wiki/Patent_troll

https://en.wikipedia.org/wiki/Microsoft_Corp._v._Motorola_In....

https://www.bloomberg.com/news/articles/2018-03-27/oracle-wi...

https://arstechnica.com/information-technology/2010/04/ibm-b...

And with all the pollutions?

You should read “Poorly made in China”. There are people and companies that do this, but, typically not accessible to small, one off projects.

The issue is, for any idea that you might have for doing business in China, there are tons of Chinese already doing that. And they know the market much better than Americans. Maybe there was a window, a few years after China became an important economy, when your dream was true. But nowadays Chinese people have much more knowledge and access to capital to do these things than American can even dream of.

Reply on news.ycombinator.com