Raspberry Pis really come with their own flavor of computer sickness, don't they.
I pulled one from the scrap pile at work and since then I've been thinking about what to do with it!
Many fun ideas, but the second I think about them in any way I realize that this can be done better and cheaper with either a microcontroller or a cheap (or often times free!) old computer.
@gigabecquerel I think one of the main niches where they make sense is when you need a unix system with GPIOs
@wolf480pl @gigabecquerel but when do you do that? Unix and GPIOs are pretty orthogonal concepts, and if you really want that POSIXness (and many people who've worked long with unixoid systems will look at you strangely if you proclaim you do and ask you which part of your application demands and hardware reality *really* reflects that need), well Zephyr on a microcontroller gets you there without booting a strange application processor with very loose control of its GPIO peripherals.
@wolf480pl @gigabecquerel … and I say that as someone who's not used non-unixoid OSes for anything non-embedded since ca 1999. The parts that make a RPi's GPIO useful are, as far as I can tell, not that it has a solid fork() syscall when running Linux, nor that you can theoretically run a postgres server with data on its SD card, nor that it's easy to port software written for HP-UX to it.
@funkylab @gigabecquerel
no, it's that it can run openssh server and flashrom
@wolf480pl @gigabecquerel SSH: well, so can larger microcontrollers. (not openssh, but I doubt you're really into Kerberos or GSSAPI and X forwarding there, but other modern SSH servers.)
Flashrom: OK, if you rather have an RPi with dupont cabling vs an EEEPROM flasher, sure, but for anything ISP-style, you'd probably want something USB-to-serial/parallel and an actual PC-style device, or go for a programmer device (say, Segger J-Link-style) rather than running on RPi.
@wolf480pl @gigabecquerel you'll also note that SSH is useless unless you use it to access/control something that your RPi is actually doing, bringing us back to the start of why you are doing it on the RPi.

@funkylab @gigabecquerel
Let's say I have an EEPROM flasher... wait no, a SPI flasher, ok.

I bought one, flashed the one NAND chip that I needed to flash, done.

A few months later, I need to, idk, do some ad-hoc interaction with some I²C device. Can my flasher do that? Or do I need to buy another device?

Now, admittedly, sth like Glasgow Interface Explorer would be a better tool here, but it's 4x more expensive than a Raspi 3B.

@gigabecquerel idk, i think old computers and rpis have a similar degree of reliability problems, just in different ways. at least rpis have energy efficiency and being silent going for them

(that said i try to use literally any other SBC instead)

@gigabecquerel Yeah, I've only bothered a couple times, when it's the relatively rare case of needing real processing power in a small package. That said, having a full dev environment onboard sure is nice!
@gigabecquerel at home we have a little Lenovo Thinkcentre, not that much more expensive used than an Rpi, doesn’t really need much more power but is a full sized PC.

@gigabecquerel Raspberry has one advantage compared to old PC - I have never seen old PC to automatically reboot and start to fully operational mode after power outage - Raspi did this many times without any problem... Second thing is a form factor and power consumption....

Plus, it was easy to connect custom hardware buttons / switches via GPIO.

But I agree in some cases much more simple MCU like Arduino can be better solution.

@gigabecquerel I have one at home and one at work each running Folding@Home and I am objectively aware of how inefficient that is but something deep inside of me is compelling me to do it. It is wasteful (of a small amount of electricity) but it would somehow -feel- more wasteful to just have both Pis sit unused in a drawer.

Per watt, I have many, many times more compute available in my (mostly idle) Mac Studio on my desk right now... and yet!