gondezee

Yea in one of my other comments down somewhere here I mentioned riscv as a hopeful future option, maybe with someone like AMD spending some resources to bring another viable solution to market. I know AMD is working on RISCV, they’ve got some low power SOCs already. I’m wondering if they got something in the pipes…

Any non-rockchip ones tho? Beagle has been dead for a bit.

Again, I’m against the IPO. Never brings good.

I don’t disagree with your points re:fairness and the community goodwill.

I think they still are the only game in town for my usecases. The support and “just works” factor is unmatched.

And with the Hock Tan era Broadcom, you’re not going to see another bcm based design upstart.

Hopefully RISC-V can ramp up and provide a viable alternative in the near term, with an AMD or something willing to provide some resources.

Be real, all electronics supply dropped to a trickle. Even if they sent them all into the channel it still would have suffered from the same dipshits scalping.

See the other hot consumer devices: ps5, nvidia 1080/2080s, Nintendo switch (even approaching 7 years old)…

Shit, I couldn’t even get an SD card for a bit.

I don’t think you understand how bad the supply chain problems got. Our vendors struggled to get fab time. When they did get wafers scheduled they had to choose which of their dies to print. When they had dies cut from the wafers, they couldn’t get packing to bond the dies into. And that’s one company. Try to align that to 10s to 100s of suppliers and their portfolios. None of those companies are focusing on their low margin devices, such as what you’d find in consumer facing devices. We had to idle our production for months across multiple product lines because of a single $0.75 chip that doesn’t have an equivalent (see above regarding ASICs). These issues finally subsided by the end of last year and have cemented a bit of patience in me that I did not have before the pandemic.

Sole sourcing of ASICs is a necessity of life unfortunately. They’re rarely pin-for-pin compatible, and hardware doesn’t allow quick spins due to regulatory recertifications any time you touch the PCB.

If your widget needs a computer in a small form factor, you can’t do much better than a pi. Not many firms have expertise of high-speed-design, so off the shelf hardware to handle that portion is pretty much a given. The fact it has exposed and broken out GPIO crowns it really as the only option in the world of SFF PCs. Arduino-class products are microcontrollers, the ATmega328P that powers them is an extremely basic MCU borrowed from industry.

For reference the least expensive industrial pc is $160 on digikey. The other SBCs are from non-reliable brands without certs, volume, etc. Rockchip based designs are going to be a no-go for many due to security implications, however imagined they may be.

These industrial customers may have started off a decade ago as a hobbiest on the pi; their volumes ultimately keep prices down for the hobbiest.

For the record I’m not for the IPO.

Working in hardware, this is a shit take. If you’re saying a hobbiest should be prioritized over keeping a paycheck coming to employees at a firm that rely on rpi parts being delivered, umm…

WITH priority as a giant customer (not with rpi, but multi billion accounts) we still were facing 72+ week lead times for components. A smaller company, more likely to use an rpi as an integral part of their widget, would be facing MUCH longer. That means manufacturing halt, and going under.

Should they have taken on industrial costumers in the first place? That’s another question. But to say my octopi server has priority over someone being able to feed their family is bullshit. Not only that but industry customers keep volumes high, allowing for lower end prices.