Eben Upton: Founder of Raspberry Pi Foundation


Eben Upton  
Founder
Raspberry Pi  Foundation 

Osborn: This week, or maybe  it was last week, I saw  a  3D printer on Kickstarter at a three hundred and fifty–dollar price point, and the guts of it are Sanguino3   and a Raspberry Pi. I think it has a five-dollar  Wi-Fi dongle for wireless networking. So people are using Raspberry Pi to build products and DIY media centers. I’m wondering if there are any really interesting  use cases that people are using the Raspberry Pi for now, which you hadn’t considered when launching this project.

Upton: Well, I think the really exciting  use cases right  this second are going to be camera use cases. We’ve got a camera out now, and that really adds a whole new dimension to it. So I think one of the ones I saw that’s been going around last week is the ZSL, which is the Zoological Society of London. They’re part of the London Zoo and they’ve got a wildlife camera project  going on. They’ve got several components. One of them is they’re going to give these cameras to kids. So all kinds of kids can build a wildlife  camera to put in their garden or on their windowsill to take a picture  of the garden at night. One really cool thing is that they’re going take these and put them in Africa. They’re going to hook them to Iridium satellite phones. That’s what they do. They put them in the field in Africa, with solar-powered batteries and the Iridium phone. They take the infrared filter off the cameras, for a sort of night vision. And they use it to keep track of animals. And they use it to look for poachers. I think the things are going to have microphones, so they can hear gunshots. If you’ve got multiples of these nearby, then you can look at the time that the gunshot sound arrives and triangulate the location of the gunshot. So you can send rangers with high accuracy to the field to find and try and track down poachers in the middle of the night. I think that sort of thing is really cool. It’s just using the Raspberry Pi as this piece of LEGO. I always thought of using it just as a computer, as just a natural  computing device. I’m a software  engineer. My natural  idea is just to write software on it. So we’ll use it as a computer, just as naturally  as you  say, people are using it like a piece of LEGO to build all this hardware.

Osborn: When it becomes the cheapest and most capable device you can buy, it’s just natural that you want to screw stuff on it, and solder things to it, and use it for a building  block.

Upton: And the multimedia stuff is already a really good one. The multimedia capabilities of the device are things like the media center. This is a device that started off as a Broadcom  graphics chip. It is not actually a general-purpose processor at all. It’s a graphics chip with a general-purpose  processor  bolted on the side. So it’s got this kind of outsized amount of graphics performance, a PS2 level  of 3D performance. It can play 1080p and high-definition  video. So I think that’s the closest use case we’ve got to my original conception of people using it as a software  media platform.  It’s this XBMC4 thing that a lot of people do.

Osborn:  So have you got a lot of Raspberry Pis out to schools and universi- ties? How is this working  as far as the original  vision? What are schools and kids doing with them?

Upton: I think it’s the early days. What we’re doing this school year is a fairly slow but steady deployment in schools led by pioneer schools. There’s always a school that has a teacher  that is particularly  keen on teaching programming. So we’re seeing those kinds of pathfinder schools doing really good work and getting through  some of the lessons that we need to learn in order to make this effective in the classroom. We have to remember this isn’t really about the classroom. This is about bedrooms. This is not supposed to be the classroom computer, because the schools already have computers.  This is supposed to be the bedroom computer. We’re starting to see parents who are doing stuff with their kids along with technically enthusiastic teachers doing stuff with their students. We’re start- ing to see a bit of it. I think what’s going to happen this coming academic year and the academic year after that is we’re going to see much more formal use of it in universities. I know a number  of universities who have already incor- porated it into their EE and CS classrooms  just because it’s a very cheap and standardized platform.  We’re going to start to see it in pathfinder schools more. Then the next academic year, the 2014 academic year, the government in the UK is revising their curriculum. It’s got a new curriculum  for computer science that has less emphasis  on PowerPoint  and Excel, and much more emphasis on computer programming.

Osborn: I wish I had something like this available when I was learning pro- gramming. I got my first PC when I was sixteen  because the price tag back then was like two grand. My family just couldn’t afford that type of expense, but I had a lot of interest in programming.

Upton: When I was a kid, I had to use old computers that I bought myself. I used to save up and buy machines. I had a BBC Micro and I had a Commodore Amiga. Each of those cost me about three hundred bucks, so about two 4XBMC  is a free and open-source  software  media player. hundred pounds. They plugged into an old TV, and those were just amazing and yet so wimpy and not powerful. I remember we got our first lot of Raspberry Pis in March of last year. We had two thousand shipped to us from China. And they came on a little pallet, because  a Pi is a little thing. The guy comes out of the back of a little DHL van and with a pallet jack, and jacks this little half-pallet of Pis into the garage. There were two thousand computers, and each of those  computers  was a hundred times more powerful than the one that I used to love  as a kid. The computer  was one of my most treasured possessions   as a kid. These things were a hundred  times  as powerful as that. That was kind of a funny moment. And then we unpacked them. We unpacked  a few of them out of the box. They all worked. It was quite something.

Osborn: You guys have done some great work, and the community  has defi- nitely taken hold of it. It sounds like everything’s  going great, but I wonder what kind of challenges you’ve had along the way, either technical challenges or challenges with manufacturing. Can you give us examples  of things that didn’t go quite as planned along the way?

Upton: We had a few. I mean, obviously, getting it down to cost was a huge challenge. Getting the billable materials down to cost was hard. And Pete Lomas, who’s my colleague, did most of the hardware  design. I think he strug- gled to get the cost of the device down. And then finding cheap manufacturing was a struggle.  We took this around to ten different UK contract manufac- turers, whom I presume are kicking themselves now, and they all just gave us these ridiculous  quotes, like ten pounds, fifteen bucks, to manufacture the thing. It was just completely  uneconomical—just  crazy money. Kind of like they would  give us a quote to just wish we’d go away, I think.  We were  very lucky because we had a guy at Broadcom, a Broadcom employee in Taiwan who heard about the project. He said, “Look, let me try to find you a Chinese  contract manufacturer.” He found me this company who makes MP3 players for Wal-Mart. That’s their main business. That was their main business before  Pi. They gave us a quote, and it was a good quote.  We had been struggling so hard to get the cost down to $35, where we wouldn’t lose money, and they came in under $35, so we could actually make some profit selling these things.

They’re based in Shenzhen, mainland  China. We had to send them chips and we had to send them money. We wired them the money, and it was probably— gosh, it was probably $25,000 or something exorbitant to their bank account. And then we had to send them the chips for the Pis. But you can’t send chips to mainland China because there’s  an import duty, so the way you get chips into China is to use a shipment  point in Hong Kong. You take them across the border from Hong Kong into Shenzhen. That’s a special transaction,  because Shenzhen is still a special economic  zone. And so we had to send these chips, $25,000 worth of chips, to these guys at their trans-shipment point. And their trans-shipment point was an apartment. It was an apartment building. We sent $25,000 worth of chips to these guys at their apartment. I was kind of like, “Ehhh, $25,000 just blown away,” you know And then there were delays, which made us worried that things had gone wrong. As usual, the guys in the manufacturing  business gave us a slightly opti- mistic quote in terms of time scale. It was supposed to take three weeks. It took eight weeks. But like I said, it was a pallet of perfectly working Pi, so that was kind of cool. So that was a bit of a challenge Another challenge, I guess, was just  from the point of managing the explosive growth. We thought we were going to build ten thousand of these things, so you can imagine  managing what happens when that turns into a billion. Needless to say, we were very lucky we had these partners. We don’t make Pis. We license the design to our partners. They make them. So we were lucky. That was a place where we would have been screwed if we tried doing this ourselves. But we have great partners, and our partners really helped us out. They helped us out so much, of course, that we were able to move the manufacturing back to the West. We moved most of the manufacturing back to the UK, where Sony builds it all on contract. Sony, who I didn’t even know did contract work. We were so pleased with the Chinese guys that we actu- ally gave them  a license to build their own Pis. So they’re now building these Red Pis for the China market, which is a nice little business.

Osborn: Cool. I’ve seen the Red Pis. I didn’t know if those were a legitimate license or not, but it’s good to know they are.

Upton: They’re only sold in China to Hong Kong/Macau. They’re completely legitimate, but they’re designed to be distinctive. They’re designed to be dif- ferent so that we can check to see that they’re not leaking into the West in significant numbers.

Osborn: So you started out with the Model  B, and then you launched that, and more recently the Model A, and now the camera. I was wondering if you could tell me what else you have in the pipeline? What’s next?

Upton: So we want to do a display board. We obviously have this other little connector that we’ve never used, which is for connecting  a display. We would really like to do that. It would  be a flat-panel  display, and I think that would be quite useful for a lot of applications. We’re obviously going to do a revision of the Model B board. We’ve always said we’re going to do that. It’s just to improve the power consumption and characteristics along with a few little minor tweaks. We thought we might do a night-vision  camera  because people  have been taking our cameras and peeling the infrared filter off with a scalpel.  This is kind of hard-core. Just the other day, we thought we might sneak in a cheek product. We’ll just buy some cameras without the IR filters for people to use. That might be a bit of fun. What else? Lots and lots and lots of software work. And lots of documenta- tion work to try to make the platform more attractive to users and trying to make it more comprehensible to total beginners. If you’ve looked at our web site, you’ll see that we’ve done some work recently. There’s a program that we call NOOBS5, which a simple out-of-the-box  expe- rience that makes it easier to get going with the Pi. You don’t need to do so much if you can’t program it yourself. You just drag files on the SD card and boot the Pi up, and the Pi does the rest. So lots and lots of little things like that to improve the user experience.

Osborn: Well thanks, Eben. I think my Pi camera is still on backorder, but I’m looking forward to getting it and seeing what I can do with it. For me it’s more about the possibilities than really having a solid  idea as to what I’m going to do with it.