- Designing a custom PCB for a more PC-like Pi, but then not merging the functionality of the official M.2 HAT despite plenty of available packaging space. Being able to just slot in an SSD would greatly expand the operating envelope of the product.
- Sticking to Micro HDMI ports, again despite plenty of space & with everyone having standard HDMI cables already on hand, or surely able to procure them more affordably.
Arguably, a non-user serviceable RPi is a bit of an odd choice, but the 400/500 are very much self-contained all-in-one products, which I guess makes it kind of reasonable.
This thing is a desktop computer. The degree to which the performance is improved by using good storage is huge. Excellent NVMe disks in the M.2 form factor are cheap and widely available. Excellent SD cards are not so easy to find. Excellent USB devices are not as good as their M.2 counterparts, are more expensive, and are sufficiently awkward dangling out the back of an RPi 400/500 that’s it’s barely worth considering.
We tend to forget that not everyone on the planet has the same resources or needs we do.
I hear this repeated over and over again in RPi threads, but nobody ever provides any proof of developing nations using RPis as affordable desktop computers apart from some PR articles every now and then involving partnerships and donations.
This feels more of an opinionated viewpoint or stereotype from clueless westerners from rich countries they have on developing countries. $50+ for a RPi might be affordable computer for you when your average computer at home is $700+, but for people in developing countries even $50 is A LOT of money and they're not gonna spend it buying RPis.
As someone originally from a developing nation, most people there use x86 PCs or laptops from the used market or rescued from e-waste imports and repaired because they're much cheaper (nearly free) and more abundant than a new RPI 5 based computer, not to mention more versatile in what SW they can run.
Or, more recently people started using old Android phones to learn on because they're also cheap and abundant and can be used for coding/tinkering. But RPis, not so much.
The OLPC project intended to provide cheap computers to developing countries also failed in part because, just like a RPI as a computer, westerners don't understand the market of developing countries and do product development from their own privileged perspective.
This isn't meant to denigrate anyone's work or effort, just wanted to share the reality in the field that I encountered.
Oh and BTW, even in the rich west where I live now, a RPi5 is pretty expresive and impractical for a general purpose computer when you can find Core i5 laptops with 8GB of RAM and 256GB SSDs for that money on the used market. And it comes with display and battery to boot and can be used on the go.
Personally I wish the world was different and those things mattered but in the reality of education it doesn't and as time goes on it feels more and more like the education manifesto of Pi feels opposite of reality. Take this like from their release:
"during the early days of the COVID pandemic, when we worked with the Raspberry Pi Foundation to deliver thousands of Raspberry Pi 4 Desktop Kits and monitors to young people studying from home in the UK"
Like what use would that have been? The systems can't run Zoom, not powerful enough to handle Google Meet, they can't handle videostreams competently enough for a conference, they don't have a microphone, don't have a camera like how would this have helped anyone learn during COVID when the main requirement was video conferencing.
The Pi is a good educational platform for e.g. Western organizations that need to kickstart education programs at a scale. It's infinitely easier to ship 100 Pis than to locally source 100 similar computers and set them up. And it's as good a platform for introducing people to computers as any, I guess. But it's not the kind of platform that gets "grassroots" adoption.
First off, if you live in a developing country, you want access to technology with wide-scale application (and the opportunities it brings). Realistically, what you want is something that runs Windows 11 or (dare dream...) macOS, that you can use to write mobile apps and games, web apps that target interesting and "hot" technologies and markets, learn system administration for contemporary platforms and so on. That alone kind of excludes anything that's not x86 or a Mac.
Hipster platforms for retrogaming and "radical simplicity" is something that engineering professionals in well-off countries can afford to do as a hobby. But if you're learning programming in the hope that it will be a development opportunity (for yourself and your family, or for your country, no matter what form it takes, whether you want to emmigrate or stay and develop your local industry) you want to learn things with real industrial relevance. The few Pis I've seen used in a scenario like these were almost exclusively embedded devices, largely because the Pi is an easily-available embedded platform, making it easy to outsource development for it.
But other than that? No, once you factor in shipping, currency conversion and all the accessories you need, a Pi is about as expensive as a hand-me-down laptop that can run a lot more software, with a lot less hassle, and isn't a fire hazard you have to be careful around with and/or fit into a makeshift case (because, if you have to scrounge money for a 50 USD Pi, even a 10 USD case is going to be kind of expensive).
Just because they're not widely used for that now doesn't make them unsuitable for that. They're an option, and options are good.
In your opinion, what would a product that successfully provides low cost computing in developing markets look like? Is one even needed?
For example, they sent hundreds of thousands of laptops to Uruguay, but they hardly got used for education because the teachers just simply didn't know how to use them or integrate them into the curriculum. But Uruguay continued the OLPC model under the Ceibal project, and it's only just now, ~15 years later, after teacher training regarding technology is far more widespread and teachers have far more experience in utilizing computers in the classroom, that it's starting to show effectiveness. And the OLPC model is considered a success in Uruguay, but it took over a decade of persistence to work itself out.
So this is actually something that RPi is tackling head on -- the Raspberry Pi Foundation is actually working hard to develop curriculums meant to address challenges in developing countries, like poor internet connectivity (which means a curriculum designed to be self-sufficient), and then also putting a lot of effort into training educators. This is before they even try sending Raspberry Pi's out, and largely is designed around using the already limited computing devices they have (not RPis) because a lot of those governments are unwilling to invest in computer education (because they have bigger problems to worry about and consider things like every student having a computer to be a luxury when nutrition is a more pressing matter on the quality of education).
Frankly, even the idea of "most people there use x86 PCs or laptops from the used market or rescued from e-waste imports and repaired because they're much cheaper" is a fair bit of a luxury and lacking understanding of the challenges of computing education in developing countries once you leave the cities and start looking at rural populations. In Mexico, nationally, 45% of households have a computer and 56% have an internet connection. Once you go to rural areas, that's ~20%. And that's before you consider socioeconomic status and the disparity in access to computers and internet. Some stats put the lower-income Mexicans at 5% computer ownership nationally.
In Rural Mexico, it's normal for a computer lab of 30-50 computers to serve a school of 500-1,000 students, or for schools to have NO computers and instead send students to nearby schools occasionally to utilize their larger computer labs for learning computer skills, and that's the only computer access they realistically have. This is really a great use case for RPi's, but like I said -- the hardware is the easy part to solve. The fact of the matter is if you sent a crate of RPi's or any other computer to these schools, they'd sit in the boxes for years because they don't have the infrastructure or the human resources to start using them. And that's actually the part that RPi Foundation is focusing on first.
True, but we must also admit that misguided prejudices about resources can lead people to miss out on some very significant opportunities, as well.
By way of anecdotal example: I have a Pi 400 on my desk that I have used as a Linux development workstation for the past couple of years - in which it has performed superlatively, I might add. Its small and light and fast enough, yet works as a perfectly good performance metric/measurement for the projects I've thrown at it.
I used it to develop a scientific instrumentation and sampling product which has gone on to produce >$40million in revenue for the customer. I've used it to develop audio capture tools which have been integrated into the r&d facilities at a major audio hardware vendor.
So, even though it might seem 'underpowered' and 'un-sexy' for most such tasks, this particular embedded developer has been able to squeeze much value out of the limits of the hardware. Not only is it a good performance reference, but it has just enough oomph to allow me to avoid using cross-compiler techniques, and in that sense is a gateway device to so many very powerful embedded Linux development tools and methods.
I really don't understand most of the comments, who cares if it just has micro hdmi ports, no nvme (whatever it is, I don't think I would care if I was 16, and I don't think I care today either) or whatever you think is missing? It costs less than $100 and runs Linux!
who cares if it just has micro hdmi ports, no nvme
Clearly lots of people. Choosing to need a special 'weird' cable, that very few people have lying around, to plug it into a monitor or choosing the storage option with the much worse IO performance and reliability just seem like strange choices, given that making the 'right' choice probably wouldn't increase the cost by all that much.
I use them quite a lot for various projects (mosty recently in my e-piano build, where the Pi 5 performs admirably pulling off low-latency audio for my piano VSTs - https://eikehein.com/piano), but while I vibe with the whole "sets a known standard, just works, network effects" advantages, there's some real annoyances that run counter to that.
For example, I'd expect the standard SBC from the Good Ones to have great upstream support, but it turns out a vanilla kernel doesn't support the Pi 5 until perhaps 6.13 due to lacklustre upstreaming efforts, meaning most alternative distros to the patched-up Raspberry Pi OS still don't support the Pi 5 properly. (I ended up managing to bake up a modified Fedora 41 because I wanted the much newer audio stack in said piano, but only with a lot of effort.)
And yes, the dreaded Micro HDMI ports, perhaps the weakest, shoddiest, most prone-to-failure port around ... why they'd put those on the keyboard one (the 500 linked here) is beyond me.
That said, I like a lot of the strategic choices they made with the RP1, and I suspect there will be more interesting HW from them coming down the pipe over time.
R-Pi offerings are fine. It is supposed to be a cheap computer, and they could get something that costs less than 200 bucks with a monitor. Definitely not bad.
The HN audience at large is still pissed that RPi chose to prioritize large industrial accounts over hobbyists during the Covid chip shortage era, and what few stock arrived at hobbyist distribution channels quickly got mopped up and re-sold by scalpers.
A few FOSS hardliners are pissed about binary blobs and a lack of in-depth documentation - which is something I do understand, after all RPi-as-an-ecosystem should be one of the larger accounts at Broadcom, so I'd expect the negotiators to finally make some progress on that front in a decade.
And others are pissed because Broadcom messed up Pi 4's PCIe implementation [1], or the RPi-side fuck-up with USB-C resistors.
And the final general group of pissed-off people I'm aware of - which again I support - are annoyed about the situation regarding the CSI and DSI (camera and display, respectively) interfaces. Cameras still are limited to official cameras only, and there's (almost?) no DSI compatible displays which means everyone has to pay HDMI licensing fees despite not needing to.
> Maybe
Maybe, maybe not. Certainly not saying they don’t but do you think that’s true about every single company by default?
My first computer had 48 KB of memory, folks don't have any idea of how much computing is wasted nowadays in their Electron garbage.
$100 for an e-waste monitor is poor value, for that money you can get a portable monitor with full sRGB, high refresh rate, etc
(It's an odd way of specifying a monitor though, usually manufacturers specify as a fraction of either Adobe RGB or Display P3.)
The product brief says 16.2M color depth, though, which would mean it's 6-bit + FRC which is bizarre to couple with an aim for DCI-P3, so perhaps not.
Quite an odd product!
I assume these are mostly all trade-ofs so you can power it on one of the Pi's USB-A ports without an additional PSU required.
Not sure why they dropped the NTSC, maybe consumers have cottoned on to NTSC keyword = bad monitor
The smaller drives are very affordable and the performance difference is huge.
The main appeal to an NVMe on the Pi isn't speed, it's cost effectiveness. SSDs are cheaper at larger capacities than SD cards. A 1tb SD card is 2x the cost of a cheaper 1tb SSD.
Storage speed is a bigger consideration on something like a gaming PC or a video editing machine, where you're going to be streaming gigabytes of textures/videos off of storage, not just trying to launch Firefox.
New Pi day is the best day!
The R.Pi 500 has 8GB RAM (phatter than the Pi 400) and a much better thermal profile than the non-kb R.Pi 5 because of the mainboard design and heatsink.
As others have said, some opportunities for the R.Pi 500 may have been missed, such as M.2 or even a backlight for the keyboard. (Project: power an LED light from GPIO!) But the TDP, performance, and 8GB RAM are excellent features. And it's a portable system with keyboard ^_^
My ideal R.Pi would be using Display Port instead of HDMI but then the SOC (which has built-in HDMI) would be a different proposition. [1]
[0] https://www.tomshardware.com/raspberry-pi/this-dazzling-rasp...
I’ve mainly been using it as a second display but plan to wall-mount it to show the family calendar, as they’ve made it really easy to flip the back leg and hang it on a hook.
I think these are going to sell really well into education and back offices.
These days, I mainly use it as a Commodore 64: https://imgur.com/Afq9uFq
Feels very much like they've given up even pretending to try and understand what their customers want. The micro hdmi thing is just incredibly stupid, they know full well how pissed off customers were over that yet did it again.
Also if you read here, their own announcement blog post, X, Bluesky and Reddit you'll find theres a fair few people sharing my views regarding micro hdmi and the nvme shenanigans.
My guess is you wildly overestimated how much people care what you care about.
Here's a much better and more in depth deep-dive into N100 vs Pi's by ExplainingComputers: https://www.youtube.com/watch?v=hekzpSH25lk
There's trade-offs in both directions, but largely rPI has marginally lower performance and power consumption compared to the N100.
If you're not touching GPIO or need to stay under 15W strictly, then the PI is the better option. N100's seem to be slightly more expensive if buying from MSRP, but the second-hand market is much more competitive and there are more discounts available with N100 systems. RPI's are considered hobbiest tech so hold their value and are rarely discounted.
You can get an N100 16GB ram with 500GB M2 storage for about $200 range, which if you were to spec out a Pi with a case, cooling, proper storage, correct cables you'd be getting close or past that.
Obviously electronics or GPIO projects I'd still go with Pi.
Also, at least until some time ago, ARM compatibility was not always guaranteed by open source projects (I think nowadays it became ubiquitous).
The ones I nabbed were the i7 6700T's with 16GB ram and a 128gb ssd. Total bargain and not at all that rare to find on there these days around that price point.
Oh they also all came with a Windows hardware license, not that I wanted it but another bonus.
- monitor can be powered from the pi 500 at 60% - has built in speakers, plus some hdmi stuff to strip the audio line to a line out on the monitor for headphones/speaker - there's gonna be a 16GB model - add the M2 slot back as a "premium" feature - the existing PCB will fit pretty nicely into a laptop case
plow all this into a laptop-like case with a touchpad and you have a ~$300-400 laptop with workable, real-world performance perfect for impulse purchases, kids, and hobbies. With market differentiators of
- low power - low cost - good software/hardware ecosystem - gpio built-in
They're really only one or two steps away from this. I'd bet by next Christmas.
Still not near Apple M1 class performance, so you have to adjust expectations for media, creative stuff. But most things I've tested work fine now on the Pi if it works in Linux at all, which is a bit nicer than last time I tried 4 years ago.
[1] https://www.jeffgeerling.com/blog/2024/pi-500-much-faster-la...
Some websites are going to perform really badly, like a lot of news websites with lots of ads, but those tend to perform bad on my gaming PC, too, and I just think if a website performs bad on a Pi 5 it's really just not worth visiting.
I have a similar gripe with iMacs and most all-in-one PCs that they cannot be used as a monitor for other devices. Once the PC/Mac inside becomes obsolete, the whole device becomes useless even though it has an excellent display that still works.
This is kind of the keyboard equivalent of this.
I tried 3 separate high quality class 10 sd cards, re-wrote all of them twice, once with dd and then with rasp imager with no success. I couldn't connect to the home router (2.4g wpa3) even after syncing the channel and couldn't even connect to my phone's open AP. But it did connect to an open Xfinity AP. It could see all the available APs, but just couldn't authenticate or connect.
I booted the rpi3 to figure out wtf was going on and it connected to everything without trouble. I then updated the pi5, but the problems persisted.
I was only able to find a few posts describing the issue, but none with a reasonable solution.
The experience pretty much killed my enthusiasm for the pi5, but I remain interested in the nature of the bug. Any thoughts?
I don't know why they did it, but it follows the same pattern as the pi400 - develop a whole new board, rather than a daughterboard for the pi4 compute module. If I had to guess, it's probably about reducing e-waste: if someone wants a compute module and it's out of stock but there are pi500s available, why wouldn't they buy a pi500, strip the compute module out and dump the rest of the machine in the bin?
And to be fair to them, they're dogfooding a fair amount - the RP1 and the RP2 are in the design, rather than a generic southbridge and keyboard controller.
That will hurt sales.
Still, I feel it is pretty reasonably priced as whole unit considering other products from them.
* no 802.11s s == mesh is best choice today. * no mechanical keyboard ? * I prefer chocolate / planck layout * fat, still too fat
options
* I need power, mobility power for computer. Why not put 2-3 x 18650 for ups or emergency working?
Not sure what the point the monitor makes.
Not to mention the GPIO header.
That said, I'm sure the included RAM + SSD isn't the greatest. The ones I got in my case (different model and store) has been working fine so far, but it's a bit of gamble.
For tasks requiring GPIO, a Zero is usually more than enough, in my experience.
N100 with 8gb ram and 256gb ssd is around 130eur on AliExpress
For people buying these as a computer the power difference isn't something that even enters their mind. How many average people do you see debating one tv over another because it uses 5w more power?
But add case, cooling, M2 board, M2 drive, power supply, mini-hdmi to hdmi and it's close.
(See other replies to my comment for links & names)
