Update On The TechCrunch Tablet: Prototype A | Hacker News Reader

Update On The TechCrunch Tablet: Prototype A

techcrunch.com

33 pointsblackswan41 comments

Browsing the comments on TC, and I found this quite telling/interesting.

http://www.techcrunch.com/2008/08/30/update-on-the-techcrunc...

In case it goes: "A couple of weeks ago I was asked to make a simple cad model for the prototype hardware. http://tctablet.altronix.se/#2.11

I would like to finish the work but since last week I have not been able to to get in touch with them on either irc or skype… :/" Refer

The idea is not bad, but they are working on obselete technology. What they should be working on are extension screens for even smaller computers - think of a very flat panel with a slot for an iphone that immediately expands the iphone screen into a 20 inch thing.

The computer itself should be expensive, but the screen part of things should be cheap.

There were rumors Apple patented this a year or so ago.

What do you mean "rumours"? Patents are open and available for all to see.

Yeah, but who actually goes to the patent office to look for ideas he can license to build products?

They are an unpleasant morass, though

Hence "patent".

Are any HN people helping out with this?

Classy. Do these guys really think you can brew up a $200 web tablet at an old kitchen table?

I've been critical of this project before, but my advice on the matter would be to refer techcrunch to the first facet of this effect:

http://en.wikipedia.org/wiki/Dunning-Kruger_effect

I don't understand how the commenters on TC think it's anywhere near an actual prototype. Groupthink?

If they're serious about this project, TC should focus on the skills they have (PR, contacts and reputation) and get this actually built by some pros rather than by some barefooted web 2.0 nerds with bits of Arrington's old macbook air.

Still willing to be proven wrong though.

>Do these guys really think you can brew up a $200 web tablet at an old kitchen table?

I'll believe it once they say what they are using for the screen...

Yeah. We don't know yet what their screen is or what it cost them. In fact, that information wouldn't be useful anyway: I'm guessing that Michael Arrington might well be able to get one model of any screen he wants for well under $200. But once he's got that screen designed into his product and is trying to contract for 2M of them... well, I'm guessing that Foxconn's sales team is pretty tough, given that they battle guys like Steve Jobs on a regular basis.

Incidentally, that's why we should be happy to see fuzzy photos of prototypes rather than more technical details. If Arrington's suppliers find out that he's committed to using Component X, the cost of producing a unit of Component X will mysteriously rise, like magic. So secrecy and a lack of public details may be a sign of good project management at work.

The question is whether the Techcrunch call for assistance has yielded a master negotiator who can eat Chinese suppliers for lunch. It might have. Silicon Valley has quite a few of them.

The screen?

It's Arrington's old macbook air screen.

http://www.techcrunch.com/2008/08/19/apple-is-flailing-badly... Quote: "That $1,800 piece of hardware has now been dismantled for parts for a project we’re working on here."

I built the prototype. I was going to afford you a response to your first comment, but you have just been ranting. You can't even recognize the difference between a 13" Macbook Air screen and the 12" WXGA screen we used.

the guy above who talked about not disclosing components as a negotiating strategy with ODM's knows what he is talking about. He is also right in that there are a few people here in the valley who really know how to bend an ODM.

Nice work. Of course I can't tell the difference, because all that anyone's seen is a photo of your prototype. I think I made a fair assumption.

Thanks for weighing in with a comment though, and as I said before, please do prove me wrong!

Do these guys really think you can brew up a $200 web tablet at an old kitchen table?

Actually, that's a very good sign. Every one of the smartest EEs I know works like this: In apartments or labs filled with piles of random wires and oscilloscopes and beat-up hard drives. You can tell which parts are the important ones because they're the ones that are only half-covered with random junk.

The first danger, when this project was announced, was that it would be a vanity project like so many others before it. It would consist of 85% talk, and 10% fundraising, and the remaining 5% would be outsourced to some team of professional ditherers with flashy Powerpoint presentations who would produce lots of shiny plastic mockups and cost overruns and never quite manage to ship anything. Or there would be a giant mob of 'contributors', each of whom is invited to bring a single resistor to be soldered into the heap.

Instead, after the initial rush we seem to have seen very little loose talk, which implies that there might just be a handful of people quietly plugging away on the thing. And now we get a benchtop prototype. All promising signs.

Of course, I have no doubt that you can produce a single, working desktop prototype of a web tablet. That's still a far cry from being able to mass-produce a reliable version for $200. But, if they want to get there, they appear to be on the right track. And, if they fail, at least they might learn exactly why, because they will have gotten into the weeds.

You can already pick up eee's and copies for $200 or so...

I'm not sure where the market is for a tablet like this personally. It's less useful than an iphone (not portable), and less useful than a laptop (No keyboard).

Less useful for what? If it had the right form factor, it might be more useful for browsing Hacker News over my lox & bagel, which is what I'm doing right now. (Commenting would require voice recognition, which is not up to snuff, yet, or a software keyboard, which would be worse. But a little Apple Wireless keyboard could be a good adjunct for that.)

And yet, this is exactly what I've been looking for for the past year or so. Not every computer needs to be a content creation device; I want a computer for my kitchen that's used for retrieving recipes. I can make the software to store and present the recipies, but I can't find a keyboard-free tablet (basically a 10" nokia 770) that would be useful for displaying the recipies in my kitchen, mounted on the wall.

One man's useless junk is another man's dream...

I love that recipes is always used as a way of explaining how useful consumer technology is. I think Apple Lisa had an ad with a women using it in the kitchen.

You'd think that recipes present our biggest problem needing a technological fix.

BTW: I agree that if it's cheap, easy & nice, there'll be all sort of uses for it.

I think this H316 tops that quite nicely: http://www.old-computers.com/museum/computer.asp?c=927

Not only does it store recipes, it has an integrated cutting board!

Well well. So now I know.

We've come a long way. From $10.6k, 70kg recipe computer with an integrated cutting board, to a $200, 0.5 kg recipe computer with touch screen.

The start-up that finally cracks this recipe problem will really make it.

I feel the same way. I've been looking for exactly this for a while now. I want one for my 3-year-old's room and I'll program some kids games and set it up to go to various kids sites.

I think this could be the computer equivalent of the Flip Video. Simply & obvious.

And that is exactly my point. It's one thing to build a prototype, and another thing entirely to be able to mass produce a web tablet for $200. And no matter how good the EE, you can't do that stuff at a messy kitchen table - you need serious resources.

Well, sure, but those serious manufacturing resources belong to OEMs in China. As a US-based designer, building a factory is not your problem -- there's no hope of doing that profitably at this rock-bottom $200 price point. Your problem is design -- figuring out which combinations of commodity parts will do what you want. After which you'll take various iterations of your prototype to meetings with OEMs like Foxconn, asking them to bid for the honor of manufacturing the product in bulk and sticking your company's label on it.

Of course, there are lots of potential pitfalls in that process -- there are several middlemen who can mark the price up, there are lots of suppliers to coordinate and play against each other, and there's the constant risk that your design is lousy, or unmanufacturable (when the factory is half a world away, it's hard to establish a good design feedback loop), or that you'll cut so many corners trying to meet your budget that your brilliant design turns out lousy. This is presumably why there are so few decent $200 web tablets.

Incidentally, though I used to be a product engineer in the cellular components industry, I'm no expert on the overall design/manufacturing process -- I was a tiny cog in the giant cellphone machine. For actual knowledge I recommend something like Bunnie Huang's blog posts about hiring folks to build the Chumby:

http://www.bunniestudios.com/blog/?cat=6

UPDATE: I just surfed through my own link, and I have to say that Bunnie is a great blogger. His hints for getting quality designs out of Chinese factories (my capsule summary: "live in China, watching the assembly line, for as much time as you and your budget can stand") are in line with my own experience.

Indeed. When I (briefly) worked on hardware, even the prototypes had to be contracted out because you can't make PCBs and solder BGA on your kitchen table. We were aiming substantially higher than a tablet, though, so maybe you can do a lot with an eval board and some sheet metal.

They've got a team of people and multiple companies involved, a basic case, hardware plan and working model, and your response is:

"you can't do that, your working model isn't good enough, you're all incompetent and overreaching, you're a bunch of nerds and you're just stealing other designer's hardware."

Why does your comment have any positive votes at all?

Much as I think the idea is silly (keyboards forever!) "Do these guys really think you can brew up a $200 web tablet at an old kitchen table?" - yes, yes they do. And good luck to them.

That's a fair synopsis of my response, yes! The very fact that they have built a prototype already means they're doin' it wrong.

The correct approach when you fly in with a set of specs (which seemed to be $200, touch screen, browser) is to find suppliers and negotiate your build cost down to sub $75 so that you can afford to bring it to market at $200

This kind of budget, with the volumes (small) and capital (small, mike arrington isn't dell) that TC would have for this project means that well before you're at the stage of ordering samples building a prototype, you would have needed to have done a whole lot of work at the negotiation table to even get into the ballpark.

I'm all for optimism, but honestly - this isn't a rails web app - it's hardware.

No, it's doable - BUT not the way they're marketing it.

As a pure gadget and not a PC, it's very possible. Take out the touch screen, add a keyboard, use ARM instead of x86, and you're all set.

I blogged about it in detail here: http://neosmart.net/blog/2008/what-the-techcrunch-tablet-sho...

A keyboard for this kind of device would be stupid. If that's what you're after, just get an Eee. And browsing the web without a touchscreen or some kind of pointing device would be a horrible experience.

They've got the concept nailed for the most part...now the challenging is building it for that price.

How do you propose that one types an email? Enters a web address? Go it iPhone style?!

And why would you need a touchscreen to browse the web? The only things you would do on a website are enter the address, scroll, type into forms, and click a link. A touchscreen only addresses the very last of these, while a keyboard deals with all of them. A button on the side of the screen - or a vertical touchpad even - would easily solve the button problem.

Agreed. This is not a laptop. It needs to be a distinct creature.

If this takes off it can have some sort of keyboard down the track. But only after the users figure out what it's for. And the keyboard is probably not central to that.

a virtual keyboard should work for most surfing...

A graph of a related effect:

http://www.phdcomics.com/comics.php?f=1056

(Same phenomenon as it plays out in academia.)

Well, a beagle board, [beagleboard.org] is only $150 bucks in volumes of one so if they can find a touch screen for 50 bucks they basically have their 200 dollar device.

Last time this came up I posted this:

Lets do it with stuff we can get and see what happens:

Main Board: (truly open, schematics and all) : http://beagleboard.org/ ($149)

Screen : Lilliput 8" 1024x768 touch screen (http://www.case-mod.com/lilliput-809gl80npct-lcd-touch-scree....) ($259)

Network : 802.11g wifi (http://www.usbgear.com/USB-Wi-Fi.html) ($30)

Battery: 7.4v LiIon 5200mAh (4 hours run) http://www.all-battery.com/index.asp?PageAction=VIEWPROD&.... ($50)

Case: Injection Mold plastic (http://www.emachineshop.com/molding_machines/injection_moldi....) ($10)

Cost to build a single one, ignoring volume discounts and cost to assemble: $498.

I still think the price is going to end up being around this. You could actually go and make this today. It would be cool if someone were to just throw up an instructable and a compact flash image of the OS for the beagle, then we could all go make our own. Someone could design a nice plastic case and have a bunch made (or just have emachineshop hang on to the molds) so they'd be cheap for all.

Your link for the lcd touchscreen is truncated with elipses. That is, the actual href="" part is truncated. Is this a feature of the hn software, or did the link get pasted in wrong?

Ah, it was from the cutting and the pasting of the actual post. Here's the link to the post.

http://news.ycombinator.com/item?id=252996

The site that sells the touch screen says "down for maintenance" right now though...

I also liked the idea of using an eeePC and using a $60 touch overlay that one can get easily on ebay. Google "touch screen eeepc". Many people have made some very nice looking modifications in this regard.

Ever hear of a project called the Apple II that a "barefoot nerd" named Steve Wozniak cooked up in his apartment on down time from his HP day job? The customers and shareholders of Apple Computer are all grateful that he didn't share your belief that "no matter how good the EE, you can't do that stuff at a messy kitchen table - you need serious resources."

Yeah, I have actually. That was 32 years ago, when you could etch boards at home and mount thru-hole DIP components in your shed. No idea if Woz etched the boards himself, but ferric chloride was always cheaper than LSD.

Sorry, but computers in 2008 are significantly more challenging to produce from scratch.

My assumptions when I made my assertion that this isn't something that can be done at a kitchen table:

- A design for a $200 touch screen web tablet is pretty much going to require a custom PCB.

- The design of that board would more than likely be a repackaged and licensed board from the cpu/chipset manufacturer, and it's not something you do in a hobbyist's workspace - you need CAD.

- You would likely not even design it yourself, you'd spec the physical, heat and electrical goals to the manufacturer and have them do it.

- Even if they were working from engineering samples and rigged up bits of gear hooked together with wire-wrap wire, proper low power and low cost components require exotic connectors and/or exotic soldering, and they just aren't available without commercial agreements.

- I assume they're going to need a processor architecture other than x86 to meet the price point (i think an Intel Atom or a Via + chipset would be approaching $50 in commercial quantities) - and I doubt anything off the shelf is going to do what they want it to.

Having said all that, some kind of mini-itx board with an atom or via processor, or a gumstix or something - and a discrete (say usb?) touch screen and lcd on an LVDS would approximate hardware-wise what they'd be likely to be building.

I have a feeling Intel would give TechCrunch the Atom processor for a ridiculous discount, just for the publicity and the proof that embedded processors can have a market.

Possibly - but I think you're overestimating the reach that techcrunch has (even within the technical sphere), and the need for publicity that Intel has.

Moreover, even if TC sold one of these to every reader and two of their friends, that's still a tiny volume in the PC market. They just couldn't possibly have the supply chain sorted out to be able to move the sort of volume that Intel would need for a good discount. Sure, they might get it stocked at Newegg, but even organic demand for a product like this doesn't generate sales. Ask Nicholas Negroponte - another greek guy!

Anyway, I think the Atom's brand recognition is probably quite good, with a sticker on every eeepc or MSI wind that's out there.

Reply on news.ycombinator.com