Technologist David Bombal has a one-hour interview with Raspberry Pi founder Eben Upton. The interview covers a range of topics, starting with the big questions about unit availability and when more stock will be available.
00:00 - Intro: Tough Environment
00:07 - Intro: Eben Upton hacked the network as a kid
00:40 - Raspberry Pi shortage (stock availability)
07:22 - People say that you're not looking after hobbyists!
10:12 - Raspberry Pi OS is backwards compatible
12:37 - The pain affecting all of us
16:33 - The origin of the Raspberry Pi // How it started
23:16 - Eben hacked the school network // Creating an environment for young hackers
32:05 - Changing the Cambridge and the World
35:00 - African growth and plans
40:03 - General purpose Computer vs iPhone vs Chromebook
43:28 - Possible IPO and Raspberry Pi Foundation
44:50 - The Raspberry Pi RP2040
48:33 - How is Raspberry Pi funded?
49:10 - How is the next product decided?
50:22 - Raspberry Pi Foundation sticking to its roots
51:17 - Advice for the youth or anyone new
56:01 - Changing roles // From tech to business
57:08 - Do you need to go to university? // Do you need degrees?
01:00:05 - Learning from experiences
01:01:44 - Creating opportunities
01:05:05 - Conclusion
No transcript is available and Eben does speak very quickly. Also published on YouTube if you do not have the obligatory LBRY account to block the algorithmic "recommendations".
Previously:
(2023) You Can Build This Raspberry Pi-Powered, 4G Linux Phone
(2023) Raspberry Pi Just Launched a Handy New $12 Tool. Here's What It Can Do
(2023) Raspberry Pi Powered Compute Blade Makes the Cut
(2023) Raspberry Pi Produced 10 Million RP2040s in 2021, More Pi Stores Likely
(2022) Raspberry Pi 5 Not Arriving in 2023 as Company Hopes for a "Recovery Year"
(2022) Raspberry Pi Adds 100,000 Units to Supply Chain, Back to Pre-Pandemic Levels in 2023
... and many more.
Related Stories
Raspberry Pi Adds 100,000 Units to Supply Chain, Back to Pre-pandemic Levels in 2023:
For Raspberry Pi enthusiasts it must seem that Christmas has come a few weeks early this year. Raspberry Pi CEO Eben Upton announced via the official blog that a 100,000 units had been secured for single-unit sales (one unit per customer) and that by the second half of 2023 it is expected that will return to pre-pandemic levels. This is good news for those who have felt the bite of the supply shortage that has dogged the Raspberry Pi for well over a year. The bad news? Raspberry Pi Zero and Zero W see a $5 price increase, but the Raspberry Pi Zero will be available for bulk purchase in 2023.
[...] In the blog post, Upton acknowledged the patience of the community and offers the 100,000 units, made up of Raspberry Pi Zero W, 3A+ and Raspberry Pi 4 2GB and 4GB for single-unit sale. We don't know the breakdown of how many of each model there will be, but Upton does indicate that it is likely that Raspberry Pi Zero W will come back into stock first. Following that, the Raspberry Pi 3A+ and then versions of the Raspberry Pi 4. Upton confirms that units are "flowing into the Approved Reseller channel now, and this is already translating into better availability figures on rpilocator."
[...] Commercial and industrial sales will be actively managed, with commercial / industrial customers set to "receive the units they need". Changes have also been made to ensure that "inventory-building behaviour which would otherwise prolong the shortage for everybody else can't take place."
To avoid cannibalizing supply for other Pi products, the next model must wait:
Few who have tried to buy a Raspberry Pi in the last year may be shocked, but Raspberry Pi's CEO has an update on the next Raspberry Pi model: it's not arriving next year.
In an interview with ExplainingComputers, Eben Upton reviews the supply pressures that have impacted the single-board computers' availability. Eighteen months into "restrained availability" of the device, Upton says the company is positioned to set aside hundreds of thousands of units for retail customers. He notes that the companies primarily taking up the existing supply of Pi units are not gigantic companies but "mom-and-pop operations" that have based their hardware products on the Pi platform and buy a few hundred Pis for their needs.
"We don't want people to get on a waiting list," Upton tells ExplainingComputuers. "We want people to wake up in the morning, want a Raspberry Pi, then get one at 9 am the next morning."
Into the near future, however, that next-day Pi is likely to be a Pi 3A+, a Pi Zero 2 W, or, later and with some luck, a Pi 4. The Pi 5 is not in the cards any time soon.
"Don't expect a Pi 5 next year... next year is a recovery year," Upton said. "On the one hand, it's kind of slowed us down. On the other hand, it slowed everything down. So there's merit, I think, in spending a year before we look at introducing anything... spending a year recovering from what just happened to all of us."
There's almost an "infinite" supply of RP2040 chips:
In a recent episode of Tom's Hardware: The Pi Cast, Raspberry Pi CEO Eben Upton revealed that 10 million RP2040 chips have been made since 2021 and that there could be more Raspberry Pi stores opening in the future.
Tom's Hardware Editor-in-Chief and The Pi Cast co-host Avram Piltch asked Upton "Why are there no shortage of RP2040 based products?" and Upton's answer "We took some big risks" lead to the revelation that Raspberry Pi purchased 500 wafers in 2021.
From a wafer, the yield is approximately 2000 die for 30mm. Newer chips, such as those in the Raspberry Pi Zero 2 W and Raspberry Pi 4 use a 45mm square die, respectively the BCM2710 and BCM2711 packages. From a wafer Raspberry Pi expect to make 1400 die.
Upton then does the math and from 500 wafers, each yielding around 21,000 die, there are around 10 million RP2040 chips.
[...] This "stockpile" of chips from 2021 are what many of us keen Pico users are currently consuming, be it in the form of Raspberry Pi Pico , Pico W or third-party boards. Upton then talks about what is "effectively an infinite supply [of RP2040]" based upon how many die can be created per wafer. This is a refreshing statement, given how global supply chains have been hit by the pandemic.
Related: Raspberry Pi Adds 100,000 Units to Supply Chain, Back to Pre-Pandemic Levels in 2023
Building Raspberry Pi clusters has never looked so good:
We've been tracking this project since mid 2021, and the time has been well spent. Ivan Kuleshov's Compute Blade is a thin PCB that packs a plethora of storage options for your Raspberry Pi Compute Module 4 (or compatible). Kuleshov's kickstarter has smashed its $522,209 funding goal, reaching $673,365 at the time of writing.
The Compute Blade is a rack-mountable carrier board for the Raspberry Pi Compute Module 4, designed for high-density clusters. The PCB is packed with features, but your eye will be drawn to the red anodized aluminum heatsink which fits over the Compute Module 4 (or compatible), providing a passive means to keep the Pi cool. This could prove useful, should you wish to overclock.
[...] The Compute Blade's strength comes in numbers, more specifically "clusters". Given the small size and blade design of the units, they will easily slide into a blade server and as long as you've got plenty of Raspberry Pi's, you'll have a powerful Arm computing cluster.
With prices starting from $65 for a Compute Blade Basic, the version we have on the bench is the $107 Dev version, which has all the bells and whistles. If you like what you see, then head over to the kickstarter page to make your pledge. [...]
Raspberry Pi gets a dedicated Debug Probe that has more uses, too:
Raspberry Pi has unveiled a new probe that allows users to debug code running on a Raspberry Pi Pico or other Arm-based microcontrollers.
The Debug Probe, which is based on the Pi Pico and its RP2040 microcontroller, is available now for $12. It's the company's first new product for 2023 and comes as the firm works to improve availability of the Zero W, 3A+, and the 2GB and 4GB variants of Raspberry Pi 4, which have been in short supply since the coronavirus pandemic.
The company decided to make the probe after noticing people were using one Pico to debug programs running on another. The probe package includes a USB to Serial Wire Debug (SWD) bridge, a generic USB serial adapter, and cables to connect to a host computer, and to the debug target.
But even if you don't want to debug code, the probe might still be a useful addition. "The Raspberry Pi Debug Probe's low price makes it a cost-effective alternative to other USB serial adapters. It has largely replaced the once-ubiquitous FTDI cable as our adapter of choice here at Pi Towers," notes Raspberry Pi chief executive Eben Upton in a blogpost.
While it has been designed with Raspberry Pi Pico, and other RP2040-based targets, in mind, he said the Raspberry Pi Debug Probe can be used to debug any Arm-based microcontroller that provides an SWD port with 3V3 I/O.
[...] The probe provides a bridge between USB and the Serial Wire Debug (SWD) protocol: on the RP2040, the SWD port provides access to the Debug Port (DP). The Debug Probe provides a bridge between USB and SWD to allow the host to access the target's debug port. Upton notes its more convenient to connect via USB, which is also the only option when using a PC or Mac.
Introducing OURphone—an open-source DIY smartphone made with our favorite SBC:
It's easy to build your own desktop PC but, if you want a phone, you usually have to settle for a sealed box that's made by one of a handful of large corporations. Maker Evan Robison wants to change all that as he posted instructions for an open-source, Raspberry Pi-powered called OURPhone with the acronym OURS standing for Open-source, Upgradable, Repairable Smartphone.
According to Robinson, the idea was to create a smartphone alternative for people looking to control their privacy. He also wanted to make a smartphone that could be easily modified and repaired, so an open-source solution was the perfect fit. The OURphone project has quite a few specs that you'll find on many smartphones including 4G LTE internet support, GPS support, Bluetooth and WiFi capability, as well as basic phone operations like the ability to call, text and save contacts in an address book.
However, instead of running on Android or iOS, the phone uses Raspberry Pi OS, the Linux-based native operating system for Raspberry Pis. This means that you have very fine control over what software you run on it, but the UI (as pictured) is not particularly touch or phone friendly. You can find all of the code used in the project (as well as detailed instructions) at GitHub.
In his build guide, Robinson is using a Raspberry Pi 3 B+ but there's no reason you couldn't upgrade it to a Pi 4. It's accompanied by a 4G HAT with GSM and GPS antennas. It has a Waveshare touchscreen for video output and user input. A camera can be attached but it isn't necessary for the build. A pair of headphones with a built-in microphone is used for call support. To keep the unit mobile, it operates off of a couple of 18650 batteries.
The housing, [m]ade out of 3mm MDF board, is a bit bulky but necessary to contain all of the hardware. It's laser cut with port access made available all around the edges.
Vlogger Jeff Geerling has an analysis of rumors of a future IPO for Raspberry Pi Trading Ltd.
But long-term, will Eben's vision for what makes Raspberry Pi change? Will there be turnover and some of the people who make the Pi a joy to use be gone?
Will the software side start leaning on subscriptions to increase revenues to make shareholders happy?
And ultimately, could Eben be replaced, and would that change things? Yes, probably, but I won't speculating about any that here. See my blog post about enshittification from last month if you wanna read more about that topic.
What I will do is answer some misconceptions I've seen about Raspberry Pi and the IPO.
The Register covered the IPO discussion the other day and while bankers have been appointed to the task, the CEO asserts that nothing will change.
"The business is in a much better place than it was last time we looked at it. We partly stopped because the markets got bad. And we partly stopped because our business became unpredictable."
"Unpredictable" is an understatement for many who attempted to acquire certain models of the computer during the supply chain crunches of recent years. "The public markets value predictability as much as they value performance," said Upton.
Previously:
(2023) Arm Acquires Minority Stake in Raspberry Pi
(2023) Eben Upton Interview on Raspberry Pi Availability Update and Painful Decisions
(2023) Raspberry Pi Produced 10 Million RP2040s in 2021, More Pi Stores Likely
(2022) 10 Years of Raspberry Pi: the $25 Computer Has Come a Long Way
(2021) Raspberry Pi Raises Price for First Time, Reintroduces 1 GB Model for $35
... and many more.
(Score: 5, Insightful) by Opportunist on Wednesday May 17 2023, @07:55AM (17 children)
This isn't toilet paper and hand sanitizer. We're not talking about a product that people will hoard because they need it, we're talking about a product people hoard to scalp it. And we're looking at a scalping price that goes well over triple the retail price (Pi4s that have a retail sales price of 50-70 bucks go for well over 250).
At that level, it pays for the scalper to throw away half of the Pis they can get their hands at and still not lose money. As long as they can continue to corner the market, they are actually quite well off keeping their prices up and sitting on piles of unsold Pis than to lower the price, because price elasticity isn't that great for these things. People who pay 250 bucks for a Pi 4 will gladly pay 150, but people who refuse to pay scalper prices will not pay 150 either, so there is zero incentive for the scalpers to lower the price. If anything, if availability increases, scalpers will just hoover up more Pis and hike the price up more because anyone paying a ridiculous 250 for a SOC will pay 300, too.
So unless we suddenly see their makers being able to flood the market with numbers that the scalpers cannot soak up fast enough, those prices will not come down.
And that does not bode well for the Raspberry Pi foundation. Until about 2 years ago, RPis were the go-to SOC for any hobbyist projects. There was no real alternative. Sure, of course there were Banana Pis and Orange Pis and a whole load of other knockoffs, but there wasn't really a compelling reason to reach for them, because all projects that you could clone and fork on github were for RPi. So people who don't really have a lot of skill adapting those projects to other brands gladly reached for RPis, because not only were they far more available than the knockoffs (something that's kinda impossible to imagine right now, right?), they also weren't really that much more expensive, usually ranging in the 10-20% range, which translates to about 10 bucks in the price segment they were in. Nothing that would warrant tinkering and fiddling with unfamiliar hardware that has shabby support by its maker (Sinovoip, I'm looking right at you!).
But when we're talking the difference between 40 bucks and 250, people will bite the bullet and start tinkering with that unfamiliar piece of crap.
And at some point, someone with the skill will fork the old Pi project and put down a Banana/Orange/Rock Pi version of it. And suddenly even the clueless person has no reason anymore to fork over hundreds to a scalper to get his hands on a Pi. There's now really an alternative.
And once that market Pi cornered is lost, it won't return. Because then others have established themselves, usually cheaper models, yes, with worse (often way, way worse) support, but I think everyone here knows that even 10 bucks cheaper will convince a lot of people to buy it, because by the time they notice that the support sucks, they already bought.
(Score: 2) by looorg on Wednesday May 17 2023, @08:46AM
This is pretty much what I hope will happen. Someone will reverse some of the older boards and make the gerbers free so you can just order your own and make your own. There are so many projects that depend on them. But they are not worth building at scalper prices. The alternatives would be to redo all those projects to use something else. But that seems to be more work involved. That said I'm not sure it will happen or can happen, I just hope that it does happen.
(Score: 5, Interesting) by canopic jug on Wednesday May 17 2023, @08:47AM (6 children)
Supplies are heading in the direction of normalization again. Eben is apparently aware of that market inertia which is why Raspberry Pi has been working so hard to keep the industrial users in the pipeline. That includes a lot of continuous negotiations about actual production needs. The interview covered that in a lot of detail. The gist is that keeping at least the minimal amount of Raspberry Pis flowing to the industrial sector ensures a greater amount of jobs retained. The hobby community is more flexible and can bounce back, at least for a while longer.
Money is not free speech. Elections should not be auctions.
(Score: 4, Interesting) by shrewdsheep on Wednesday May 17 2023, @09:56AM (1 child)
The best move therefore would be to introduce a new model, or maybe just announce it. This would substantially devalue scalper's stockpile, forcing their hand to unload. The announcement of a gap year is really detrimental, IMO.
(Score: 4, Interesting) by aafcac on Wednesday May 17 2023, @12:59PM
Or people could have some patience and just preorder their Pis and just wait. That's what I did when I wanted my Pi 4. It was a bit of a wait, but ultimately, buying from scalpers just encourages more scalping. But if you either just wait or buy from one of the other Pi-like SBCs, the upside to scalping drops a lot. But, I do think that it probably will improve in the future as production gets back to normal after years of pandemic and then all the international sanctions being put into place against various countries.
(Score: 4, Interesting) by JoeMerchant on Wednesday May 17 2023, @10:08AM (3 children)
As a hobbyist, I bought a Particle Argon for a project that needed low power. Then the Pi Pico came out and I abandoned the Particle because it didn't seem to have the long term availability of the Pico, and sure enough they discontinued my model not long after, never did provide full software support for all the features of the board, etc. Meanwhile Pi Pico goes in the opposite direction, and the lower price is a bonus.
For a single project, I don't mind a higher price point, the labor far outweighs price considerations. But Pis are so much less expensive than most viable competition that I stock multiple spares, which reduces physical damage anxiety that comes with prototyping hardware, and means that spares are laying around for future projects when I get the time. The Particle would have been a one off but the Pi Pico has 3 working widgets and plans for 3 more when I get the time.
I would really love to get my hands on more Pi Zero 2Ws and 4s, the shortage sucks that way, but cross model compatibility and the half dozen Pis I have already mean that the Pi is still my platform of choice even with the ongoing supply issues.
🌻🌻 [google.com]
(Score: 2) by Freeman on Wednesday May 17 2023, @01:42PM (2 children)
I got my hands on a Pi4 a good while back and still haven't taken the time to do anything with it.
My hopes is that I'll be able to buy some version of a Pi that will eventually be an acceptable desktop replacement for my Wife. The Pi4 is just a tad slow when it comes to the likes of YouTube, etc. Even moderate browsing is a bit of a chore on the Pi4. I believe I got the 4GB version. At this point, I really wish I'd ponied up for the 8GB version. As part of the issue is 4GB is just not quite enough for a tab hoarder. The 8GB would be much closer, but it's still just a tad slow. Would be awesome, if the Pi5 allowed for native m.2 boot and/or NVMe support. So that you don't have to stuck using a SD card for boot or some extra work to get USB booting to function. Even just getting a functional USB 3.0 as native boot would be "okay".
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(Score: 2) by JoeMerchant on Wednesday May 17 2023, @03:49PM
Pi4 is an acceptable media server, running LibreElec.
It's not a horrible desktop.
I have one running a custom mp3 juke box player which is a pretty killer app - like having your own radio station but better, but it just corrupted its 2nd microSD card so my next Pi project is putting that app on an external SSD ($14 for 120GB these days and the USB3 adapter is a bit less.)
The thing I don't like much about the Pi 4 is that it can overheat if you aren't careful. But, then, LibreElec on a Pi Zero (slow interface but solid playback) also can overheat too...
🌻🌻 [google.com]
(Score: 3, Interesting) by GloomMower on Wednesday May 17 2023, @08:41PM
As LVM pointed out, a better desktop replacement might be a https://www.bee-link.com/ [bee-link.com] Some are new for as low as $135.
(Score: 3, Informative) by GloomMower on Wednesday May 17 2023, @01:01PM (3 children)
> And at some point, someone with the skill will fork the old Pi project and put down a Banana/Orange/Rock Pi version of it. And suddenly even the clueless person has no reason anymore to fork over hundreds to a scalper to get his hands on a Pi. There's now really an alternative.
I've tried to use Orange Pi's and Banana Pi's. There are many different models with different networking and graphics chips. They are not consistent in themselves and the drivers are bad. What use is the Orange Pi's version of the Zero W2 when the wifi drivers are very broken? It needs more than someone to port the project, the kernel drivers need overhauled and fixed.
Armbian is a step in the right direction, but for a lot of them, you download some linux image from a forum that links to a google drive.
(Score: 1) by stack on Wednesday May 17 2023, @03:58PM (1 child)
This is partly why Rocky Linux has made a big investment into a lot of the popular alternative SBC's. They are actively looking for people to assist... :-)
(Score: 2) by Rich on Wednesday May 17 2023, @04:32PM
I would have assumed from its name and the context that this is geared towards Rockchip SBCs. Yet it's just a RHEL clone named in favour of CentOS co-founder Rocky McGaugh. (And I'd rather have it if an SBC-supporting Linux stayed with dpkg, much like everyone else, and was very careful with RedHat's great "innovations").
(Score: 2) by Opportunist on Friday May 19 2023, @08:23PM
The Banana Pi Zero is fairly decent once you know how it works, but don't expect ANY support for it. You will spend some time figuring out what's going on, and Google is more of an aid than the manufacturer.
But then again, it's 30 bucks instead of 90 (provided you can get your hands on a RPi Zero in the first place).
(Score: 2) by VLM on Wednesday May 17 2023, @04:53PM (3 children)
There is a higher level market for pi replacements above the $50 boards, I have a couple beelink "$175" mini-PCs. About twice the size of a Pi, 8 G ram, quarter TB flash (don't have to buy a SD card) HDMI and wifi work on Ubuntu out of the box, 2 GHz Intel processor, runs on 12 volts at a couple amps (far less than one most of the time, but it's "peaky" up to 3 A occasionally)
Beelink sells about a zillion different models the one's I bought last winter are I'm sure obsolete with better specs available now.
I mean, yeah, I could spend more money and have a larger pain in the ass by trying to set up a 3 host K8S cluster (using RKE2 for whatever that matters) AND have all the PITA of ARM software instead of intel OS installation, but why bother if the mini-PCs with small SSDs and in a case sell with power supply for $75 less than a Pi4 without storage and without case and without power supply?
Now when you could get Pi for $25 that would be worth some PITA to set it up, but when the intel based competitor has higher performance and sells for cheaper...
Pi's just kind of dead in the market. If it takes "real software" spend less money on a mini-PC. If it is a glorified "blink a LED using an arduino" then get a $5 ESP32 dev board and install circuitpython or micropython (choose your poison, LOL) on it.
(Score: 2) by canopic jug on Wednesday May 17 2023, @05:12PM (1 child)
Beelink sells about a zillion different models the one's I bought last winter are I'm sure obsolete with better specs available now.
The Raspberry Pi OS supports even the oldest models. That's something which can't be easy to hammer out every time the upstream distro, Debian, gets a kernel or other upgrade. However, it is both beneficial to end users and important strategically for the project that they spend the resources to keep it that way.
Money is not free speech. Elections should not be auctions.
(Score: 1, Insightful) by Anonymous Coward on Wednesday May 17 2023, @06:54PM
x86 and Intel/AMD integrated graphics are better supported by Linux than any weird ARM device, RPi included.
(Score: 2) by GloomMower on Wednesday May 17 2023, @06:41PM
Depends on the use case, big difference in size and power usage, heat dissipation, compared to the Pi Zero W2.
But thanks for sharing, the beelink's look pretty cool for certain use-cases.
(Score: 2) by takyon on Friday May 19 2023, @03:43PM
People who are mad at RPi and shortages will come crawling back, guaranteed.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by ElizabethGreene on Wednesday May 17 2023, @09:12PM
My takeaway was they are expecting all products to be fully available by the last quarter of the year, with products becoming available more-or-less in the order they were originally released.
I want a 02w to Octoprint-ify my new 3d printer, and I'm giddy to hear that.
(Score: 2, Informative) by bmimatt on Wednesday May 17 2023, @10:36PM (14 children)
- It's a tough environment. It's been particularly tough environment, obviously, for the hobbyist community, which are really at the, kind of, at the heart of what we do at Raspberry Pi. I think the good news is you look at what, what we've tried to build with the foundation in terms of infrastructure, and it's to build the organizational infrastructure, the social infrastructure that you require in order to have a good experience with computing.So, you know, what did the social infrastructure look like for me as a child? Teachers who'd give me respond to be hacking the school network by giving me an account on the, on my own private account, rather than sending me to the headmaster and putting me in detention, right. You know, people who would engage with well-intentioned, but some potentially disruptive enthusiasm in a positive rather than negative way.- Hey everyone, it's David Bombal back with a very, very special guest. Eben, welcome. - Good to be here. - I wanna talk about the whole history, about like how this stuff came about, you know, from your ideas to reality. But before we get there, I think a pressing question a lot of people will have is, I can't buy it.- Yeah, we are in this very, very strange time. You know, we are still in the aftermath of the pandemic. I guess an interesting thing happened about three years. Well, obviously an interesting thing happened three years ago, but from the point of view of people who want to build electronic products, the interesting thing that happened in the spring of 2020 was that there was a, I guess an assumption, that the pandemic was gonna lead to a worldwide recession.A major worldwide recession. That led to electronic product manufacturers and manufacturers of electronic components decided to run their inventory down, you know, the last thing you wanna be stuck with in a recession is a big part of unsold product. And so there was a period of time, probably three to six months, where people were destocking, or running down their inventories.And at the same time that that was happening, the recession that people were predicting wasn't arriving. And the reason for that, obviously was governments took kind of furlough type action to support the income of people who weren't able to work 'cause of the pandemic. And so you're in an environment where you have some people who are stuck at home with nothing to do because they're on furlough.You have some people who are at home and have work to do, or at home and need to study from home. And all of those people are more, not less inclined to buy electronic products. And it took a little while for all of us who make electronics to realize this and to adjust our purchasing behavior accordingly. And so probably by the time you get to the autumn of 2020, you're in an environment where everyone's run their inventory down and suddenly everyone wakes up one day and says, you know, I don't have enough stuff.I'd better go order some more chips. In the first instance, that leads to a fairly small shortage environment, you know, lead times start to go out. But as we saw, I guess a little earlier with, you know, toilet paper and toilet paper and hand sanitizer. Once there's a perception of shortage, people change their buying behavior.And so you get these kind of panic dynamics. You get people trying to build inventory. And we've seen now a two year, really a two year, two and a half year period where it's been hard to get your hands on electronic components on any kind of sensible lead time at any kind of sensible price. That's affected Raspberry Pi.It's affected a lot of people. You see these pictures from Germany of these enormous car parks, outside car factories full of cars that have no engine These stories about people putting it, having one engine management computer and you put it in your Volkswagen, you drive it out into the parking lot. You take the engine management computer out and you put it in the next one and you drive that out and then you sort of start a stockpile almost finished cars.This has been affecting a lot of people and it's affected Raspberry Pi. The interesting thing about Raspberry Pi for the, you know, we've been selling them for a little over 11 years now, and for most of that time, the model we've had for getting your hands on Raspberry Pis, we'll put, well probably maybe be half a million Raspberry Pis out there in Chandler at any given time.And whether you're an individual who wants to buy a single unit, or an industrial customer who wakes up one day and needs 10,000, 20,000, 50,000 for a product, you can pull that from channel without really disturbing the inventory that's out there. We've gone from that world to a world where, you know, they are in shortage.You know, where we're having, you know, where hobbyists having to join queues, you know, either having to look very, monitor very closely to see inventory going out to the channel or join queues at resellers.
(Score: 2) by bmimatt on Wednesday May 17 2023, @10:39PM (11 children)
But it is, it is a problem and it's the, I think I've said it's the hardest decision I've had to make in 12 years of running Raspberry Pi. - I've heard you say, you know, it's really painful for you. - It's incredibly painful, right? 'Cause I'm a hobbyist, right? I mean, I'm in this for the hobby side.I'm not in this as, I'm not OEM, I'm a hobbyist, right. I think what we gotta remember is when we talk about OEM we talk about industry, you're not talking about billion dollar companies. There are billion dollar companies using Raspberry Pis, deploying Raspberry Pis at scale. We have a few of these kind of whales, you know, people who are buying one hundreds of thousands of Raspberry Pis a year.But by and large, when we talk about, you know, I've said we have two, two and a half, three thousand OEMs, that kind of implies your average size of OEM customer is, you know, on the order of a thousand units, yeah. These are, a lot of these are mom and pop, organizations, small businesses. These people who won't make the mortgage if they can't get Raspberry Pis.At the point where somebody has designed a Raspberry Pi into a product, there is a moral obligation to try to supply them with material. You know, that balance, striking that balance between supplying OEMs, keeping OEMs alive and supplying individuals. It's an incredibly difficult balance to strike. I think probably, you know, one place where you see the consequence of us trying to make these decisions is in this sort of differential return to availability that you're seeing, you know, something like Zero or 3A+that doesn't have a substantial OEM footprint. Those are the devices that we see. So we're doing our best and what our best looks like is bringing the products that don't have a substantial OEM footprint back into general consumer availability first and then back filling with a product like Raspberry Pi 4, and Raspberry Pi 4, and compute module 4, which both consume from the same, 'cause this is all about core silicone.You see these FIFOs of core silicone coming in. So, you know, 2835 for Pi 1 and Pi Zero. 2837 are various sorts for Pi 3 and 3+ and compute module 3 and then 2711, and those are all 14 nanometer chips. And then 2711, which is a 28 nanometer chip. You see that, that's a different FIFO. And so you see, you know, these FIFOs, they have different incoming rates and they have different levels of OEM demand pulling on them.The 2711 based products, the PI 4 and CM 4 were such a massive and instant success in industrial space that those are the ones that are gonna take the longest to come back into general availability. - I've heard you say this and I think people maybe don't know about it or forget about it, that the software that runs on the latest Pi 4, that can still run on all the versions, right? - Yeah, that's right.And, you know, the enormous amount of work goes in here into maintaining every single way back to, we have, we have the boards, which called alpha boards, which are they kind of don't look like a Raspberry Pi. They're so old, they have another, a completely different form factor and they're from the summer of 2011.The software, and there are about 50 of those in existence and we have between 5 and 10 of them here in this building. And there are people in the wild who have those and will complain if a new release doesn't run on this 256 meg alpha board from the summer of 2011. Yeah, we put an enormous amount of work into it.What that does mean is yeah, if you need the, if you need the raw computing power of a Pi 4, then it's hard times at the moment. But you know, if you need a machine that can run the platform and you are prepared to put up with lower levels of performance, we are already, and it's April, we're already in a position where there is very, very good availability of some of those devices that run the software stack.- That's very different to other companies where you think you have to get the latest to be able to use all the product. So what you're saying is even if I get like a Pi Zero, I can do stuff at home, I can learn, I don't need to get the latest and greatest, the most expensive product. - Yeah, and we're working as hard as we can to bring the latest greatest.You know, we love Pi 4. You know, we're putting, just as we put enormous amount of effort into maintaining compatibility with the older devices, we put enormous amount of software effort into between major generations of Raspberry Pi. You always see an increase in the performance of the latest and greatest device that tends to be driven by software optimization.So, you know, that is the latest and greatest device. It's probably the place where most software optimization effort is going, you know, here at Pi House. And so of course, you know, of course we wanna get that into people's hands, you know, but we've all the way through this situation, you know, we've tried to find mitigations, we've tried to find ways that people can stay engaged with the platform.We've continued to produce new accessories. Yeah. We launched a (indistinct) of camera products this year. And yeah, I think there is a school of thought that says, why are you producing these camera products and you can't buy Raspberry Pis? Well lots of people have for Raspberry Pis. To give people interest in new hobbyist experiences, to give people interest in new educational experiences.You don't necessarily have to ship a new pi. You know, you can ship some sort of new accessory, which just lets people get a little bit more out of the hardware they already own. - Yeah, I feel for you because, I mean, I think you're in an impossible situation, because I think you've mentioned it in previous interviews I've watched your sales were like flatlining rather than what you were hoping.- If you look at our annual sales figures, six million in 2019, seven million in 2020, unconstrained, weirder but unconstrained, right? Seven million in 2021. Very, very substantial constraint conditions. So if you imagine that year, we went into that year with half a million units of customer backlog, left the year with four and a half million units of customer backlog.Now you can't just add those numbers together. You can't just add seven and four and say, that would've been an 11 million in a year because there's the toilet paper, the toilet paper dynamics are in there, you know? Double ordering, triple ordering, ordering for buffer, ordering for buffer build.But it's clear that there was an increase in demand. I think we actually, we may have seen an increase in demand as other platforms became less available, as other embedded modular platforms became less available, you saw people migrating to Raspberry Pi, not just 'cause they like the platform, but out of the perception that it had good availability.So you sort of see an increase in demand in 2021 that in inability to grow the business, in order to meet that demand and then last year five million units, right. So you see an actual decline for the first time ever in Raspberry Pi history, a decline in our production numbers. - 7 Million to 5 million, that's quite a drop.- 7 Million to 5 million And you can sort of see that if, if you think that maybe 60, 70% of your demand is OEM industrial demand, then, and you and you attempt to fill that demand that explains why there's not been a huge amount of volume left over to serve our, you know, our original core market. It's incredibly painful.These are incredibly painful decisions. You know, these are not, you know, I don't kind of not throwing myself on anyone's sympathy, but they are, they're not trivial decisions, you know, and they cause pain to people who, about whom we care enormously. And so, you know, you go looking for mitigations, you try to do what you can, you try to find pools of inventory that can be allocated.
(Score: 2) by bmimatt on Wednesday May 17 2023, @10:41PM (10 children)
You try to find ways to, you know, we have a wonderful network of approved resellers. And of course I feel for our approved resellers, 'cause that's their businesses as well, you know? Some of our approved resellers have both an industrial and a consumer practice. And those people have been a little bit insulated from these challenges.Some of them are very, very consumer focused. And those people, you know, if we're not giving them stock to sell, they can't run their businesses. And again, many of those are mom-and-pop businesses, right? So you've got these kind of balance that you're trying to find pockets of inventory that you can allocate into consumer, trying to find, you're working with your reseller partners to try to find ways to avoid scalping.You know, if a unit that goes into the hand... - It's a big problem, yeah. - Yeah, absolutely, right. And you know, we've largely got in front of it now, but you know, there was a period of time where you did have people who were buying units and then reselling them on eBay. And that's just, that's painful.That's just painful to us, you know. We don't wanna see someone having to pay $100 for a $50 product, $150 for a $50 product, you know, that's not the vision, right. So yeah, a lot of our resellers have got extremely creative in the way that they've tried to avoid this kind of behavior, Adafruit for example.You know, you gotta, they do two-factor authentication effectively. You know, you can buy a Raspberry Pi per cellphone number. You know, these sorts of things where you find something that is expensive to have lots of, cell phone numbers, and then you tie sales to that object, you know, your driving license.I mean, you know, present your driving license and you can buy one, you know, one in video graphics card per quarter, per driving license. So, you know, you get people coming up with these creative solutions. We think we have got in front of it, but it's, you know, I don't think we'd ever attempt to deny the pain and the difficulty that this has caused people.- It's caused, I mean it's affecting all of us. You didn't start this to, you know, just try and make money. - Yeah. - You started this for a good reason. So let's perhaps talk about that a bit. - Yeah, let's do that. - Like how did it start? Because it's kind of changed, you know, the original vision has changed a bit over time, 'cause you're talking about OEMs and stuff like that.But perhaps you can take us back in time. - Let's take it all the way back in time. All the way back to the 1980s. So I was a kid in the 1980s and I had a BBC micro computer. So I had this another great Cambridge designed, You know, the CEO, look back to the days of Ed Zack of Morris Wilkes designing, you know, building Ed Zack at the university in the late forties, early fifties.People have always built computers in Cambridge. And there are two kind of great, I mean there were a number of them in the eighties, but of course two huge, hugely successful and influential ones. The BBC Micro, which is made by a company called Acorn, which then went on to become Arm. And then you have the Sinclair Spectrum, which was produced by Sinclair Research.You go all the way back and in the corner of every classroom remember it, and there was this beige box and you turn it on and it goes boop, it makes this lovely little two-tone kind of beeping sound gives you a programming prompt. And I got my introduction to computing through one of these machines. That's where if you walk out into the office here and you talk to engineers, the vast majority of people are vaguely my age, between their thirties and their fifties, will have got their first experience of computing,their first experience of engineering using some sort of platform like that. And really what Raspberry Pi is, Raspberry Pi has its origins in a recognition in about 2005 that the disappearance of those machines, those machines kind of disappeared from, became non-current in the early 1990s. That a delayed action effect of that, the disappearance of those machines, was a decline in interest in computing among young people.And that was feeding into university recruitment. It was feeding into industrial recruitment. It had all sorts of negative consequences, and negative consequences for the country, of course. I also believe it had negative consequences for individuals. One of the really, really interesting things about those machines was that you were finding people, I'm very, you know I always be very focused on Cambridge, 'cause it's somewhere I love and I have this great relationship with the university, but actually one of the interesting things about these machines was it was providing a gateway into good, stable, white collar employment for people who wouldn't have considered themselves to be particularly academic, you know, people who would sit down in front of their spectrum and suddenly go, hang on a second, you know, I can write a computer game and I can sell this, you know, I can burn onto a cassette and I can sell it to people for money and I can go work for a games company.I can leave school at 16 and I can, you know, spend my life in a really great white collar job, right? So they're all of these advantages to the general availability of general purpose computing in people's homes, which was a historical accident. Yeah, these machines weren't in people's homes because there was some grand design, there was maybe a little bit with the BBC micro, but there wasn't really a grand design to promote computer literacy.It was just that it happened.
(Score: 2) by bmimatt on Wednesday May 17 2023, @10:44PM (9 children)
The cheapest object, the most convenient object you could build that would do the things people wanted to do. Play computer games, do their schoolwork, you know, do their taxes. The cheapest object you could build to do that was the general purpose programmable computer. And so you kind of got all of these side effects from the general existence of these machines, both for individuals and for institutions and for the country.These machines disappear, all these nice side effects go away. And really Raspberry Pi starting in about 2006 through to 2008 when we incorporated the foundation, was kind of hunting for a way to build a platform, originally at a very small scale, to build a platform that we could get into the hands of young people to reboot some of that enthusiasm.- I love what you said there, because I mean, you know, people who perhaps are really good at programming are not perhaps good, like in a typical white collar type job. But I mean, computing gives you so many options. - Yeah, it's wonderful. And I mean, I've had a very, I've had a very traditional, very academic route, I guess through the first 25 to 30 years of my life.So I went to school, I studied, I specialized in science in the sixth form, 16 to 18. I went to work for IBM for a year. I was very, very lucky that IBM had a wonderful pre university program. A chap called Ian (indistinct), who ran their pre-university program and kind of get 50 to 60 18 year olds every year.Give them a year at IBM doing real work. And the wonderful thing about Ian was that he'd recruit people. He'd have some in his group in Warwick, in the Midlands in the UK, and then he'd farm them out to other IBM locations in the UK. I was very lucky to be in his group, but you know, of the 60, say 50 of them wouldn't have been in his group, but God help you if you took one of Ian's pre university employees and made them make the tea, or do photocopying.He had to give them real work to do. And I used to sit near his office and I used to hear him shouting, every now and then he'd be on the phone to somebody shouting at them because he discovered that a PUE had been betrayed and had been made to make the coffee. And so I did that, came to Cambridge, did my undergraduate, dropped outta my undergraduate.I actually run a games company, founded a games company when I was a third year here. Did that, sold that, did a PhD. So kind of very traditional academic flow through. But kind of in parallel with the academic track, that kind of very craftsman like thing, you know, I guess started the games company. I've always been kind of interested in the craft of computing, the craft of computer programming, the craft of software development.There are some people who only have the academic side and there are some people who only have the craft side. I mean, as I say, I've met some amazing engineers, software engineers. Left school at 16, never felt they needed to do anything else. They just went straight into a programming job, and thrived in it.And so you get some people who have both sides. But I think, you know, what's worked well for me and what we tried to promote with Raspberry Pi, and what I think giving academically minded young people access to a general purpose computer gives is the opportunity to have both sides, is the opportunity to have both the, and it doesn't really matter which order you learn them in.I learnt the craft, I was writing assembly language when I was 14. I learnt the craft and then I learnt the theory. But you get people who learn the theory and then the craft. But if you don't have some sort of platform that you can hack around on and a low consequence environment, if we talk about Raspberry Pi, you know, this says you've got a Raspberry Pi Zero there, it's a $5, it was, $10 now, but it was a $5 machine.Machines that are so cheap and, you know, that are fully featured Unix machines, but are so cheap that if you break it you can just go and get another one. You know, giving people that kind of low consequence hacking environment is a really important contributor to building these rounded, both to giving the people who don't consider themselves to be academic, a route in and to give the people who do consider themselves to be academic, to have kind of a more rounded approach to what computing really means.- You mentioned the word hacking and a lot of people that watch my channel are into hacking of various types. I believe there's a story you said when you were at school where you, you know where I'm going? - I do, I do. - You can tell the story. This is about me having the school with 12 classes and 13 accounts.
(Score: 2) by bmimatt on Wednesday May 17 2023, @10:55PM (8 children)
There's there's a series of machines called, there's a company called Research Machines. They still exist. And they built a machine called the Nimbus, which was actually a slightly horrible machine. And it was one of these nearly IBM compatible machines. It's sort of an Intel 8086, 8088 based machine that was almost IBM compatible.And we had one of these with an ethernet 10BASE-T ethernet running around this. Every class had an account and I somehow managed to jailbreak my way out of the class, out of the sandbox that the classes were in to get myself access, particularly to basic. I mean on any of these machines, the goal was get yourself access either to a DOS prompt, or to a basic prompt.- So they would force you into a sandbox, like you could only use certain applications, is that right? And then you'd break out... - Yeah, that's it, that's it. And then usually, you know, usually it was a matter of finding something, you know, finding something you could write to the disk or finding a moment, a crack, you know, a moment where hitting the right key would drop you out.Usually when you were launching an application, you know, if I hit the right key here, then it'll drop me to you, something will crash and it'll drop me to a prompt. And I got, yeah, I managed to... In both my middle school and my secondary school, I was able to find ways to get out the sandbox. And of course the way you get paid for getting out the sandbox and for not teaching other children how to get outta the sandbox, is you get your own account.You get your own account with no sandbox. And so both my middle school and my secondary school, I had privilege network access. It's interesting because what's really important, you know, as much as, you know, you look at what we've tried to build with the foundation in terms of infrastructure, you know, it's very easy to, I mean particularly 'cause we're at Raspberry Pi limited here, we're the trading subsidiary of the foundation.It's very easy to focus on the physical objects, or the, you know, the hardware objects, or the software objects. But a lot of what we've actually tried to build with the Raspberry Pi organization is to build the organizational infrastructure, the social infrastructure. To rebuild the social infrastructure that you require in order to have a good experience with computing.So, you know, what did the social infrastructure look like for me as a child? teachers who'd give me respond to be hacking the school network by giving me an account, my own private account, rather than sending me to the headmaster and putting me in detention, right. You know, people who would engage with, well-intentioned, but some potentially disruptive enthusiasm in a positive rather than a negative way.You know, I talk a lot about, there's a fellow called Alan Drew, who sadly is no longer with us. He ran IT for Leeds Permanent Building Society. You know, one of the large financial institutions in Leeds when I grew up in Yorkshire. And every other Friday evening, he would come around to my house in his MPV and I would've packed up my BBC micro in its expandable polystyrene case.And he would drive me to the local college and we'd go to computer club from 7:30 to 9:30 in the evening every other Friday. And that was just a wonderful environment for me, because it gave me a venue to see what other people, it gave me a venue to show off. It gave me a venue to show what I'd made.And it gave me a venue to see what other people had made and be inspired. You know, be encouraged, be inspired, be just eyes to things I don't know whether I hadn't done a good thing. And so you kind of look at what the foundation's been doing, you know, how does that reflect in the foundation's activities? Well, a big focus on clubs, Code Club and Coda Dojo, the kind of two big club programs and then the teacher training programs, PiAcademy, and then the National Center for Computing Education.Yeah, the foundation's been very focused on generating both the informal out of school or after school environments, but also ensuring that teachers are well equipped to be enthusiastic. Do not feel threatened by a child who comes and does something wacky in class. You actually feel that they've got the support, feel they've got the knowledge, feel they've got the collateral behind them required to deliver, you know, deliver a quality computer learning experience to everybody.- I love that though, because your teachers didn't, like you say, send you to the headmaster. - Yeah. - They kinda like grew that passion that you had. They didn't try and lock it away. - Mr. Stewart and Mr. Wright. - Goodness to them because I mean, if they had like come on top of you like a ton of bricks, we might not have had this today.- That's it, right. You know, everyone can put their, everyone who's had a positive trajectory in computing can put their finger on one or two moments where they might have had a negative experience, or just a neutral experience. You have this learning curve that has 10 print I am the best 20 go to 10 at one end of it and has, I'm a professional computer scientist, chip designer, software engineer at the other end of it.And a lot of the work that we do is about sanding that learning curve. It's about taking the bumps out, you know, and providing people with the underpinnings. 'Cause everyone's gonna hit over you, you're never gonna sand it completely smooth.
(Score: 2) by bmimatt on Wednesday May 17 2023, @10:57PM (7 children)
Everyone is gonna hit some kind of road bump at some point.And the question is, what infrastructure do you have to support you, and what infrastructure do you have, you know? I grew up in a lovely town. My father was an academic, you know, it was, I had a lot of privilege as a child, you know, so I don't wanna talk about too much about privilege, but I had a very, a wonderful privileged, blessed childhood.But a lot of people don't have that. And so you'll find if you grow up in Cambridge, your school will be besieged by volunteers to run afterschool clubs. If you're 30 miles north of here up in Norfolk, maybe you don't, maybe your school has nobody who is interested in running an afterschool club, you know, you maybe have no volunteers in the community.That's one of the reasons why engagement with teachers is so important. I come from this kind of anarchist kind of approach to this where, you know, everything I did was like, let's hack the school network. Let's go to a club. Let's go to an afterschool club and then go to the pub afterwards.You know, I come from that kind of world. And when I came into Raspberry Pi, that was the ethos that I brought with me. And probably, you know, we talk about the various ways in which our thinking has matured. And one of the ways is the idea that that is an incredibly, that is the product of an incredibly privileged background.To think that that is a viable strategy for all children, everywhere. And actually that engagement with the formal education system. The formal education system is the thing that gives people the underpinning, is the thing that gives people the backstop, the guarantee that they will get some exposure to computing.If you go and look at the computer industry, we hark back to the 1980s and it was a wonderful time. But if you go and look at the diversity of people in the computing industry, middle-aged people in the community industry, it's all people who look like me. It's all people, you know, it's white men from middle class backgrounds.Why is it white men from middle class backgrounds? Because those were the people who were messing around with computers in the 1980s and 1990s. You know, by engaging with the formal system, yeah, that's the only way that you're gonna reach the people. That's the only way that you're not gonna continue to perpetuate the demographic weirdness of our industry.And yeah, it's great. Yeah, I hope we always have lots of like, you know, people who look like me in this industry. But I hope that we also, well the, one of the interesting things when you look at the people who are participating in Code Club, for example, which is our the foundation's program for 9 to 13 year olds, it looks a lot more like society.You know, over 40% of the people going to Code Club are girls at a critical age, 9 to 13, at a critical age where girls often disengage a little bit from, you know, society. Something in the culture tends to, at a point where girls are often performing as well, or better than boys, something in the culture starts to tell girls that this is not for them.If you are a school which has a large number of deprived children, if you measure deprivation by the proportion of people who are entitled to free school meals, subsidized school meals. If you're a school with a high index of deprivation, you are slightly more likely to have a Code Club than if you're a school with a low level of deprivation.What that means is, if we kind of look 20, 30, 40 years into the future, there's real opportunity, there's real chance that if you were to walk out into an office, into a computing office, that the people you would see would be in all dimensions, much more representative of society as a whole. And that's gotta be a good thing, right? These are amazing opportunities and we have to make sure that, yeah, the two angles, one for individuals, they're amazing opportunities and we have to make surethat people have access to those opportunities. Two, we're an industry that's massively hungry for talent and we can't afford to let some pool of talent sit idle, sit fallow, you know, we can't afford to not discover that great computer programmer who happens to be from some underrepresented background.- I love that, I mean, the story, if I understand it correctly originally was you were at Cambridge and you noticed like the student applicants were going down and Cambridge is obviously very different to like, people perhaps in Africa. You've been in Africa. - Yeah, yeah. - What you've created has fulfilled that.'Cause they've, you can tell the story obviously much better than me, but like the applicant numbers went up, but you've also affected the whole world. - Yeah, it's amazing how moderate the vision was, I suppose, you know, that we were... - To what its become. - Yeah, yeah, it's, you know, we thought well, we'll build a thousand units and we'll get them into the hands of the right thousand kids, and then if a 10th of those apply to study computer science at Cambridge.We went from, we wanna find about a hundred kids a year to come to Cambridge to study computing. And we were down to about 200 applicants. I mean, that's incredible, right?
(Score: 2) by bmimatt on Wednesday May 17 2023, @10:59PM (6 children)
In 2008, if you wanted to get your kid into Cambridge to read physics, the best way to get 'em in was to get 'em to apply to computer science and then switch once they're here, right.You know, there aren't many courses at Cambridge that have a two to one application ratio, right? And so this is a really simple ambition, you know, give out a thousand computers, get an extra a hundred applicants. Oh, now we've got 300 applicants for a hundred places. That's really a material change.Now what's actually happened is devices gone out in vastly larger numbers than that and they've gone out in a vastly more disorganized way. It's not that kind of like rifle shop that kind of like, so let's go snipe that population there. Let's go get these people. It's very broad, very broad deployment of units.Only a small number of which got into education, but that small fraction is still vastly more than a thousand units. Last year we had about 1500 applicants. It's gone from being one of the easiest courses to get into to being absolutely one of the most challenging courses. People should still apply. We do still need talent, but it is certainly very challenging.And we've seen that. And that's not just us, that's not just us, that's all of these other people who woke up at about the same time and realized there was a problem and tried to solve it. All the people who started afterschool clubs, all the organizations that started trying to train teachers.The government's curriculum reform, I mean we have an incredibly better curriculum. You compare the curriculum, the very office skills centric, ICT curriculum, that we had back in 2012 that was then disapplied and replaced by this completely insane, well designed computer science curriculum from about 2015 onwards.So you see all of these things have happened and they've all had a contribution. But the net, the aggregate effect, Cambridge has been measurable in this kind of, sort of sevenfold increase in the number of applicants between 2008 and about 2019. And you see that mirrored everywhere else. You see that mirrored elsewhere in the UK, you see that kind of resurgence elsewhere in the world.Raspberry Pi I think has been an important part of it. But the interesting thing is that our ambitions have grown as we've succeeded more, we've got hungrier, we haven't got less hungry, we've got hungrier. And so the question is, you know, can we do, okay, we've done it for Cambridge, can we do it for other institutions in the UK? We've done it in the UK, can we do it in other developed world countries? We've done it in the developed world.Can we do it in low and middle income countries? You know, one of the things I'm absolutely most excited about at the moment is our impact in Africa. I've had a chance over the last six months to go to Ghana, to go to Kenya. I'm just back from Morocco. And you just see this level of desire, this level of hunger for the opportunities that general purposes computing brings, the economic opportunities, the educational opportunities, the chance for advancement for the individual, the chance for advancement for a country.We went to, very lucky, went to the, in Morocco, went to the University of Mohammed VI, Polytechnic, about halfway between Marrakesh and Casablanca. This huge, these vast buildings, this enormous technology university that's been built over the last 10 years really on the greenfield side in Morocco. And just that ambition.You talk to the people there and you see that there is the ambition not to catch up with what we have here in the UK, what people have in North America, what people have elsewhere in Western Europe. But to leapfrog over it, you know, a real desire to kinda seize control of their sort of technological destiny, and to do that not just for one country, not just for Morocco, but to share that advancement with all the countries in North Africa and in sub-Saharan Africa.From day one we've always had a very strong reseller presence in South Africa. And sort of, for a long time, Africa for us really meant South Africa. I think it's the case for a lot of technology companies, right? Africa really means, maybe South Africa, maybe some of the countries in North Africa, maybe Egypt, but, you know, certainly other bits of francophone North Africa and sub-Saharan Africa not really figuring on Mike Buffon who's our chief commercial officer.He said, I went to Kenya with him, October last year and we did a panel and he said, yeah, he's had roles at a lot of, before Raspberry Pi, roles in a lot of electronics companies, distribution companies for 30 years, all of which have global in their title.
(Score: 2) by bmimatt on Wednesday May 17 2023, @11:01PM (5 children)
And the first time he went to Africa was in January last year.- It's crazy. - It's incredible, right? - Yeah, and there's so much talent and young talent. - Absolutely. - So many young people. - It's a young continent, right? - Yeah. - It's a young continent. A couple years ago, just before the pandemic, but we hired a chap called Kenan Colo, who's based in Lagos.He was introduced to us. We were very lucky that we have, so John Lazar, who's the chairman of the foundation is a South African originally, has a real focus on kind of, I guess entrepreneurial engagement in Africa. And he was able to introduce us to Ken. Ken joined us as our strategic partnership director in Africa.And really what he's been doing since he joined us is actually not, some of the aspects of how the market works are different in Africa, but really what he's building is infrastructure, which feels very familiar to us. What he's doing is he's building reseller networks. The thing that's powered the adoption of Raspberry Pi in our kind of existing core markets is this network of about a hundred approved resellers who we, you know, we regulate their pricing.You know, we require them to sell at no more than a certain price. We provide them with access to our products at a favorable price. We provide them with clickstream, you know, so if you come to our website and click on buy for a product, we will geolocate you and send you to an approved reseller in your local geography.So the approved reseller's kind of the heart of the Raspberry Pi, kind of outbound operation. We haven't had those in Africa. And what we've been doing is country by country booting up very familiar looking AR networks. The way that those ARs do business with their customers is often different from the way that you would do that in the UK, the way that we get product to them is often different.There are often some logistical challenges, taxation challenges, duty challenges of often getting hardware to them. But what they're doing is very familiar, what they're doing. They are just approved resellers. We were in Morocco in March, onboarded and approved, onboarded an approved reseller who looks like they were about to become our largest approved reseller, kind of almost from a standing start to become our largest approved reseller outside in Africa, outside of South Africa.People who've been actually selling Raspberry Pis for a long period of time, but not as an approved reseller and that's, non-approved resellers is fine. There's nothing wrong with non-approved resellers. It just means these are not people where we, this is somewhere where we don't regulate the buying experience.And so you take somebody who's been actually a passionate advocate for Raspberry Pi for a number of years, but outside the tent, you bring them into the tent and you give them access to supply and you give them access to pricing and all of a sudden these things, they explode, and they're exploding because they're a little bit like Raspberry Pi in the UK in 2012.There is latent demand for these products in the African market. And all we're doing is we are dropping well-developed in obviously products which are vastly more sophisticated than our 2012 era product. We're dropping these products into a, like putting a crystal in a supersaturated solution. You drop them in bang, you know, suddenly you're accessing this enormous reservoir of demand.It's amazing. - I mean we, the the computer industry needs more and more people. I mean... - Oh yeah. - And what you're doing is you're empowering all of these people to learn. I wanted to ask you, you use this term a few times today, and I've heard you use it before, general purpose computing versus say an iPhone.What's the difference and why is this so important? - We are really passionate advocates for general purpose computing, so a general purpose computer is a (indistinct) machine, right? It's a thing that can do anything. You can program it to do anything. You can sit in front of it, and you can program it to do anything a computer can do.- Are you, you were thinking people would use it for games and then you use this example of cucumbers or something like that. - Yeah, that's it. And that's the lovely thing about general purpose computing is how, you unlock creativity when you put general, if you put special purpose computers out in the world, all they can be used for is the thing that you intended them for.So you put a games console out in the world, it gets used to play computer games. You put a, you know, a mobile phone out in the world. It gets used to, yeah, make phone calls, run apps, but fundamentally run apps that other people have developed somewhere else.
(Score: 2) by bmimatt on Wednesday May 17 2023, @11:05PM (4 children)
You know, play content that other people have created, not create really meaningful content yourself.When you make an appliance, you bound the things that people do with it. The point about general purpose computer is you can do literally anything. So, you know, when I conceived a Raspberry Pi, I didn't wanna put the GPIO header on there. I didn't see what the point was of the GPIO header. And Pete Lomas, one of my co-founders designed the Raspberry Pi 1 and he said, "Well, you've got these GPIO pins on the chip, we should bring them to a header." And I'm like, "Are you sure? You know, I'm not sure how robust those pins are, you know? What if someone sends static shocks into them with their finger?" And sorry, you know, I wasn't designing it. So I yielded to his enthusiasm for it, And of course, yeah, thank goodness. I mean, if I hadn't then, you know, my vision that this was, that people were gonna use, you know, I suffered from a paucity of vision.I thought that because I use my BBC and (indistinct) to write computer games, that's what people would do with Raspberry Pis. And it turns out the majority of what people in education do with Raspberry Pis is physical computing. They do robotics, they do sensing, they do monitoring, they do those sorts of things.And that's what gets people excited. So you see it's the power of general purpose computing. General purpose computing doesn't have to exist. As I, I think I said earlier, there was a period of time when the general purpose computer just happened. If you wanted to play computer games, the cheapest thing you could make and give to somebody to do that happened to be a general purpose computer.And then over time, people found ways to make both technical and business model innovations, enabled the rise probably of, you know, certainly the 16 and 32 bit games consoles were a combination of technical and business innovation, which kind of nibbled away at a big fraction of the deployed base of general purpose computers.And there's really no reason why that nibbling won't continue until there are no general purpose computers. I mean, the Chromebook is a fascinating example of this. You know, what was the remaining citadel really of general purpose computing until Raspberry Pi came along. The reigning citadel of general purpose computing in the consumer world was the PC and the Mac.And those were largely laptops. And that felt like a very secure citadel. But of course the Chromebook is a really interesting attack on that. People have managed to figure out a way to make an object that looks a lot like a laptop, can do most of the things that consumers want from a laptop. And yet it's not really a general purpose computer.And so Raspberry Pi, you know, we will always make general purpose computers. There is no point in us making appliances. That's not what we're on the planet for. We've sold 50 million of them, right? We put a lot of general purpose computers into the wild and we're just passionate advocates for it on all fronts.Both as an educational tool, but also as a platform for OEM customers, for industrial customers to innovate around. - Eben, I've heard rumors, and I think you've said it on other interviews about perhaps a possible IPO? - Yeah, I mean we, you know, I have this piece of paper I reach from time to time that says we look for when people ask that question, you know, we always look at ways of funding the future growth of our company.We looked at the possibility of doing an IPO, we invested a little bit of money. I think you can see this from our accounts. We invested a little bit of money in I guess 2020 calendar, '21 in understanding whether an IPO might be a track we could go down. the semiconductor shortage in two problems with doing an IPO for the business.Certainly in that sort of timeframe. One, the shortage makes it hard to be predictable. It makes it hard to understand what the business is gonna do and what does the market want. The market wants predictability. And the other thing is just generally there is this notion the markets are closed. By this time last year the markets were closed.There was a huge amount of IPO activity in 2021. And then there was really no IPO in London, there was no IPO activity really in 2022. So that's not really something that we have been able to pursue. I mean, I'm not saying it's something we won't come back to. I mean, it's a potential path in particular, it's a potential path for the foundation to realize some of the value in the trading company in Raspberry Pi Limited.So it's something, I mean, it's an interesting, it was an interesting journey, you know, investigating. It was an interesting journey, we've met some great people, had some great conversations, and it is something that, you know, it might be something we'll come back to in a few years. - You've developed your own custom silicone, is that right? - Yeah, that's right.
(Score: 2) by bmimatt on Wednesday May 17 2023, @11:13PM (3 children)
So 2021 January, 2021, we released Raspberry Pi Pico, which is a lovely look, it's a lovely product. But the thing that's really interesting about it is it's the first time we've released a product where the core silicone was also designed at Raspberry Pi. So we talk about it being a board that has two Raspberry Pi logos on it.It has a logo on the board and it has a logo on the chip. Now, that turned up at a really interesting time for us. It turned up just as the supply chain challenges were starting to bite in our core business. We talked about mitigations, we talked about trying to find our enthusiast community, other things to be enthusiastic about, other things to be involved with during this kind of difficult period, being able to offer a platform in PICO and RP 2040, which we knew we could make as many as we want, you know.There was no problem if you want 10 million Picos or 10 million RP 2040s, I can probably deliver them to you pretty quickly. There's no real constraint on my doing that. So having a long constrained platform during this difficult time was nice. RP 2040 is a lovely chip. Our team here did an amazing job designing it.It's very nearly perfect in a way that, you know, an application processor, the big chips that go into the big products can never be perfect. They're too large, they're too complicated to ever be kind of perfectible, you know, they're extremely good, but they're not perfect. That chip is within its kind of design envelope, is pretty much as good as it could possibly be.And it's very kind of satisfying to have been involved with and to be shipping a product based on a piece of silicon that's just, just right. You know, something that's based on the other team here, they've been working with microcontrollers for decades, individually, obviously within the organization.Raspberry Pi's been shipping microcontroller base products for a decade, you know, the core, not the core product, but the accessory products almost all have microcontrollers of some sort. Individuals have been involved with microcontrollers for decades. Everyone accumulates kind of ideas of what's good, what's bad, you know, what features did I enjoy in this microcontroller, what features do I wish I'd had? Really our RP 2040 is kind of a, I guess a consolidation of all those insights that members of the team have had over the years.All them consolidated down into this piece of silicone, obviously a piece of silicon that's on a fairly advanced process node by microcontroller standards on 40 nanometers. We get about 21,000 die out of each 300 millimeter wafer. One of the reasons why it's in good supply, get 25 wafers, it's half a million chips.Kind of the minimum unit of wafers that you'll get from a foundry is generally a cartridge 25 wafers is half a million chips. And so even fairly small supplies of upstream wafer starts, turns into a lot of inventory for us. - But you mean that raises the issue, like how do you fund that? I'm assuming it costs a lot of money to design all of that and rather than like just use merchant silicon.- Yeah, I mean RP 2040 I think was about my mental model of it. And it's always interesting to look in the accounts and see whether your mental model aligns with the reality is that it was a about a $5 million, - Wow. - About a $5 million program. - And you sell 'em for? - We sell 'em for 50 cents.- Wow. - Right? - It's a lot of investment. and we don't make 50 cents on each unit. It's not a thing that cost 10 cents to make. So you sell it for 50 cents, you wholesale it 50 cents, retail it between 70 cents and a dollar depending on volume. So you, you know, you gotta sell, if you wanted to make your money back on chips alone, you'd have to sell 20, 30, 40, 50 million chips.Now the interesting thing of course is we are in a privileged position because we also make Pico, we make the product that's built on the chip, and of course the margins on that, again, that's a $4 product for Pico, but the margins on Pico are substantially better than the margins on RP 2040. So you can kind of say, well look, I'm gonna pay this program back on the back of Pico, which I couldn't have done without it.And then I'm not gonna feel so bad about the fact that I'm only making 10 cents on an RP 2040. - How do you fund that? Is IPO an option? Is like private? Do you have private investors? - It's a little bit of a mix. I mean, historically Raspberry Pi is entirely funded by, it was organically funded.- Just from sales? Yeah, just from sales, just from profits on selling Raspberry Pis and accessories. So, you know, over the years, you know, we've returned between 30 and 40 million pounds to the foundation, to our shareholder, but we retain some of the profits we make inside our organization as well. And that's the money that's paid for things, programs, like Raspberry Pi 4, you know, Raspberry Pi 4 was not a cheap program.Pretty much like RP 2040 and Pico.
(Score: 2) by bmimatt on Wednesday May 17 2023, @11:19PM (2 children)
What we'll fund whatever we, when we decide what we're gonna do next, it's what we'll fund what we do next. - How do you decide what to make? Like you've been through so many different like, products, like, we've got some of them here, and some of them we don't.But how do you decide what to make? - We make what we wanna make, we make what we want ourselves. It's an unfashionable way of designing products, I guess. These days you're supposed to focus group everything to death. The nice thing about making a product that you want, that's another fruit company that's famous for doing this as well, right? - Yep.- If you make products that you want, then at least one person wants them. If you make products by focus grouping and by interpolating, the thing that I think is a killer in product design is where you have two, you identify two genuine pools of demand and then you think, wow, it'd be efficient if I could make one product that serves both these pools of demand.So you would interpolate and you make some products in the middle that lands between the two and these of no interest to either side. So you could, it's very easy if you do that kind of approach to make a product that has no customers. If you make a product you want at least there's one customer. And in practice there's a lot of people like me, there's a lot of people like the, oh, I'm not the, I, it's not like I sit down and design these products myself.We all sit down together and decide what we wanna build. There's a lot of people like us and that's why there's demand for Raspberry Pi 'cause we are an engineer-led organization that is designing products that are appealing to engineers. - I love that because I think a lot of people, especially the techy people watching will say that companies lose their way when it's all about sales.And it's no longer the engineers. - Yeah. And it doesn't always hit. I mean we, you know, we've had some products. TV hat I think was not the biggest success in the world. I mean, it's sold out. You know, we have the, you know, we always fairly conservative when, particularly when we do something which is novel, we're always fairly conservative about how many we make.And we do always trade out eventually. But we have a few products that took a few years to trade out the initial build volume. I think TV hat was the obvious example. And we have, you know, a recent, a good example of recent product, Global Shutter, the global shutter camera. I mean, it's an amazing product, but again, it's a product that has a very, it could be quite nichey.- I was gonna say it's niche product. - Yeah, it's gonna be super interesting to a niche. And we dunno how big that niche is. And so we've been quite conservative about how many of those we built. Even though we really believe in it as a product, we have to not get ourselves in a situation where we sitting on a million dollars of stock, it's gonna take us 20 years to sell, right? You've walked a journey and I mean, you've shared some of like the products that you make are the products that you'd want.But like you've walked this road now and there's a lot of people watching that are perhaps starting their journey, or you know, they're inspired by you. Many people are inspired by you. What advice would you give your younger self, or someone who's starting out, you know? We've spoken about some of this offline and...- I have a general reservation about advice, right? The general reservation is that a lot of what's happened, a lot of what's happened to anyone who is successful in business is they've been lucky. And you can stack the deck and you can make your own luck. But in the end, you know, I look at the history of Raspberry Pi and so many of the things which were really, really meaningful to us were very contingent.You know, they didn't have to happen. It was somebody bumped into a guy, you know, Alan Mycroft one of our founding trustees bumped into Pete Lomas, another founding trustee who designed the first product, bumped into him at some conference at Imperial and took him for a walk in Hyde Park and told about what we were trying to do and recruited him that way.So, you know, you have those, you have those kind of, they're often personal relationships kind of happen stance. So, it's an important caveat. And anyone who stands up and says, I made it in business 'cause I'm a goddamn genius, is a liar, it's just, is a liar, is delusional. Yeah, I mean, you can stack the deck.I mean, things that have been good for us, getting onto things that have been good for me, getting in touch with the market early. People talk about minimum viable products. Building minimum viable products. It's easy to sit in an ivory tower and think about what the perfect product would be and spend years working on it.When really if you got in touch with the market, well there's nothing to stop you keeping working after you've launched that first product, you can keep working. You can keep fettling it, but you'll be fettling it in the context of having market feedback. And that's been very good to us. I do have an MBA and I'm quite an advocate for doing an MBA at some point in your career.It's a little bit like I'm an advocate for hacking on computers when you are eight years old and then doing a computer science degree. I'm kind of an advocate for founding a bunch of businesses and screwing them up and then to some degree and then doing an MBA. I'm much more of an advocate for that than I am doing a computer science degree if you've never put under a computer, or doing an MBA if you've never started a business, right.I found, I did the MBA here at Cambridge and I found there were a lot of really useful strategic insights, particularly about, you know, how you structure a business to be sustainable in the long run. How you manage talent, how you decide what work you're gonna do inside the organization and what work you're gonna do outside the organization.Companies are so ubiquitous in the world that people don't think very much about why companies exist at all. They're like, why can't we have a world which is just built out of free contractors, freely contracting entities that do commercial transactions with each other. And of course, you know, you have this tension, but that approach doesn't work, because you have transaction cost economics, right.You know, it's very labor intensive to constantly set up and tear down commercial relationships between freely contracting individuals. So to a certain size, corporations have an advantage because they, you know, they hide all that transaction cost economics. You just get a pool of labor, get a pool of talent, and then apply it to problems without having to constantly write new contracts and in above a certain size.Because inside a corporation you're effectively, it's a planned economy, right. And we all know planned economies don't work very well, right. On the one hand, why don't you have a whole sea of freely contracting individuals, on the other side, why don't you just have megacorp? Why don't you have, well, so union was megacorp, wasn't it? That didn't work very well.So there's a dynamic tension there. And so sort of understanding what you do inside your organization and what you do through contracting, what you do through third parties, understanding what sorts of innovation can be supported in a contracting fashion and what sorts of innovation really are best done inside the planned economy of a single corporation is an important, it's an interesting, I mean, just like with computer science, right? There is a body of professional knowledge.There's a body of theoretical practice associated with running a business and it's really easy to disdain it, right, you know, because, we have this kind of culture of the macho, you know, guy who started off, you know, Alan Sugar kind of character, you know, guys who started off running a, you know, selling stuff on a stall.Ends up running a big corporation, right. And never did a day's executive ed in his life. And that's great. And there are savants like that out there. These, the lucky people or savants out there who will kind of walk that road. But for most of us, I think that's an expensive way sort of talked about, you know, the cost of an MBA, you know, tens of thousands of pounds.An MBA is the cheap way to get business experience. The expensive way to get business experience is to start some companies, screw them up, learn the lessons. It might save you, you know, those MBA fees might save you and save a couple of companies the experience of being founded and screwed up. - You started like hacking when you were young, then you went in like hardcore tech, you were designing chips, really technical, and then like your role's kinda changed over the years.Now you're more business focused. - Yeah, I'm extremely business focused now. I try to keep up with the people we have. We have an amazing workforce here, obviously. Engineering workforce here is by far the highest average capability level that I've ever engaged with. And you know, I go out there and I have to try and help people understand what to do, think about what to do next.And that's kind of an intimidating environment.
(Score: 2) by bmimatt on Wednesday May 17 2023, @11:22PM (1 child)
I try to stay current. I am not especially current. I live in fear of becoming less current. I'm still at a level of technical ability, a level of technical involvement where I can at least ask at least half the questions aren't dumb, which is good.That's a good ratio. If I can keep the ratio above 50%, then, you know, you sort of still feel like you're hanging on by your fingernails to the edge of technical relevance. And it's a great bunch. It's a great bunch of people. I mean, it's a wonderful group of people and of course, many of whom I've known for well over a decade now.It's a wonderful group of people to be doing innovation with. - Here's a nasty question, alright? Do I need to go to university? - Oh, do you need to go? Do you need to go to university? I don't think you need to go to university. I think there are some challenges. Society's got itself into a bit of a hole in respect of whether you need to have a university degree.And there's a risk that what's happening to degrees is they're becoming badges of conformance and badges of conformity. And that what you are saying is, I was prepared, here's a token, I was prepared to spend 50,000 pounds to buy this token. And what that token tells you is that I conform and what employers want, what a lot of employers want is people who will...All employers, I mean, it's a respectable want for an employer to want employees who conform to certain standards of socialization. And the real risk is that all we've done is we've created this token. And if you don't have the token, then people think, oh, you must be a bit odd. Perhaps you'll be a bit, you're a bit odd.Do I really wanna hire someone who's a bit odd? That's really dangerous place for society to have got to. So I think there's enormous value in degrees, and I don't just think that there's value in degrees, which are, I mean, my father was an English academic. I'm not one of these people who thinks that university is a training ground for employees, you know, and the only worthwhile degrees are the ones that couple directly into some transferable commercial employable skill.You know, it's a horrible way of thinking about education. And so I think there is enormous value in it. And in the particular case of computer science, I think it has value in the sense of a little bit like the MBA thing. I once met a guy who was working for a graphics company in Finland now Fins are amazing, a lot of amazing Finnish engineers.And this guy was working on a compiler for GL, the GL shading language component of the, the OpenGL driver that we were gonna license for the video call platform. And he had never been to university, amazing engineer. And he had written this compiler from scratch, and he had, through intellectual brute force, had invented, you know, all of compiler technology all the way through to the mid 1970s, you know, from scratch.He could have done a six week compiler course at university and not had to do that. And it would've got him to the state of the art in 2000, not the state of the art in 1975. So there are, it's important again, not to disdain the body of professional practice, and, you know, having respect for the body of professional knowledge can sit alongside having respect for and acknowledging the importance of practical endeavor as a way of kind of levering that.If you wanna go get a job in, you know, as an engineer in the community industry, yeah, it's good to get a computer science degree. At some point it'll save you effort in the same way it will save someone who wants to start an MBA will save a business person some effort. But it's not to say you can't get there the other way, but I honestly think it's a harder road.- I mean, I love what you said. You're talking from experience. Like you had some businesses, some of them didn't go so well, some of them went well. If you'd done the MBA, it's like learning from people who've walked that road before you. - I certainly, in terms of, you sort of think about going back to university.So I did my first startup when I was a third year here at Cambridge. I dropped out of the fourth year of the engineering program to run that startup, found a couple of things. One, I think like a lot of engineer led startups, we had a big pile of paper in the corner of the room. The government would send us paperwork and we'd put in the corner and ignore it, or our bank would send us paperwork and we were, you know, it was a successful business, put in the corner we'd ignore the pile of paper.And after a while we were actually based at London Business School in London and in the end we basically ran out into the corridor and grabbed a passing MBA graduate by the collar and said, if you come in here and sought down that pile of paper, we'll give you some stock. And that certainly instilled in me a very great desire to go and get an MBA, so I didn't have to do that again.You kind of have these, you have these experiences running businesses, which kind of make you alive to the stuff you don't know. So that's the kind of business example for my first startup. The technical example is I didn't have a computer science degree. I had lot of, I'd done a lot of hacking and I did an engineering degree, a physics and engineering degree here at Cambridge.And I was recruiting people into that business. Now I was just aware that they knew stuff that I didn't know.
(Score: 2) by bmimatt on Wednesday May 17 2023, @11:25PM
You know, data structures and algorithms type stuff, compiler type stuff, digital electronics type stuff. And I was just aware that there was, there was clearly a missing chunk of stuff that I needed.And that's why I went back and studied the diploma here. That's why I came back and did the conversion course. That's what kind of drove me out of my first startup and back into academia. - I mean, this is a great place to work, or if I wanna go and work for No, it's not that, well I think the name and from what I've seen is amazing - We do fun stuff, right? I mean. - Yeah, exactly, I mean.- And it's the deepest of deep tech. There's not much tech this deep in the UK. - Exactly. - You know, there's a lot of, and you know, that's not to disparage the FinTech. There are people who are building writing software, you know, running consumer focused software. But, you know, there aren't many companies doing the super deep stuff here in the UK.- But how would I get here? - So what do we hire? They generally do have pretty good academics, but I wouldn't say that we hire, you know, we're not particularly Cambridge snobs. I would say we have as many people from York as we do from Cambridge. There are other universities in the UK compared to Cambridge and Cambridge produces some amazing graduates.But if you think about that, that trade off between practical and theoretical skills, there are other trade off points and there are for our needs, there are the universities that hit those trade off points in their courses a little better than Cambridge does. So when we're taking people from Cambridge, we're usually taking people who happen to have gone to Cambridge, but who were hobbyist programmers beforehand and who've got their hobbyist stamp, their crafts person stamp somewhere else.So we take people with good academics, but not necessarily, doesn't necessarily have to have to be Oxbridge. We only really take people, certainly into the hardware and software teams. We sometimes talk about have you ever written a computer program you didn't have to write, you know, one that you didn't have to do for your course, one that you didn't have to do for your employer.Did you ever build an electronic product or write a computer program you didn't have to write. So people who have demonstrated some kind of personal passion for this stuff, it's a pretty good discriminate at that, right? You know, that really does discriminate very well. And, of course, you know, one of the challenges is the over discriminates a little bit because what you're doing is you're discriminating against people who've had, who've lacked the opportunity to do that.And so to some extent what you're seeing with Raspberry Pi, it's a closed loop system, right? it's a, we kind of soup to nuts. You can see we're a foundation which is dedicated to giving every child the opportunity to discover that they like competing at a venue in our hardware platforms and our educational resources.A venue in which they can work on that. And that then makes it more justifiable. Will in due course make it more justifiable for us to operate this hiring policy, which could be seen otherwise as being discriminating in the bad sense, not discriminating in the good sense. - I see it a lot in like other fields as well, like technical fields.If you aren't prepared to put in your own effort and do your own thing, it's like, are you really actually, is this really what you wanna do? And I mean what you've got with Raspberry Pis, I mean obviously it doesn't apply to everyone, but for a lot of us, we have no excuse because we've got a device that we can hack on, or - Yeah, that's right.Play on or do something with. - And one of the wonderful things here is we now employ people, small number of people in their twenties who got their start on the Raspberry Pi. - That's great. Right, so that's, that closed loop system is now starting to establish itself and we've got more work to do to make it, to keep it going, to make it happen.But we're on the road now, which is, and earlier back to the idea that the computer industry, the workforce and the computer industry in 10, 20, 30 years time is gonna look a lot more like society than it does today. And that's gonna be in part, I hope because of the work that we've done with the foundation.- Eben, I really wanna thank you, you know, for taking so much time to talk to me, just for everyone who's watching, we've done a lot today and I really wanna thank you for, you know, taking all this time to share, but also to, you know, encourage millions around the world to change their lives through what you've created, but also sharing your knowledge.- Thank you. I mean it's weird, right? I've been working on this since 2006, right. You know, this is a, it's getting on for a 20 year endeavor now. And the remarkable thing about it is it's still good fun. You know, I'm an inveterate. Before we did this, I saw myself as an inveterate, doer of things for four years.I thought that was my time period for working on anything was four years. This is coming up to towards 20 now. So just wonderful to have a chance to chat about it. Thank you very much. - It's fantastic. And I mean, I just wanna say thank you for changing lives through what you've created. I mean, I'm from South Africa.You know, these kind of devices give people in places that are perhaps not as privileged, the opportunities to change their lives. So thanks so much. - Awesome. Thank you.
(Score: 2) by bmimatt on Thursday May 18 2023, @10:04PM (1 child)
TFA: "...No transcript is available and Eben does speak very quickly...."
So I drop a GPT transcript as a public service and get modded Troll?
LOL
This place is funny sometimes.
(Score: 0) by Anonymous Coward on Friday May 19 2023, @02:54AM
> a GPT transcript as a public service
Those words don't go together. You posted a bunch of words. There is nothing which ensures any level of accuracy behind the claim that it is a transcript. Rather it is the opposite is more likely since the words were of the alleged transcript were hallucinated by GPT.