4 ms·
I miss RPi a great deal. I recall that a core piece of RPi's strategy was to use excess capacity to create an awesomely capable and super inexpensive product f
by WhatsTheBigIdea 4y ago
I miss RPi a great deal.
I recall that a core piece of RPi's strategy was to use excess capacity to create an awesomely capable and super inexpensive product family. That core piece of thinking was instrumental to RPi's success.
Semiconductor production capacity was already getting tighter before Covid, but Covid fully shifted the space from capacity excess to capacity drought.
The same leverage that super charged RPi's growth is what is now snuffing it out.
I really want to see RPi come back. It seems that will only happen if...
1) Some dramatic shift in global dynamics returns the semi conductor industry to the capacity excess state.
or
2) RPi manages to extract itself from the business/leverage model it used to power its growth and find a new way to do its thing.
Both seem pretty unlikely in the foreseeable future.
BTW, the Beaglebone has a completely different mission, purpose, and philosophy. It's a pretty excellent piece of hardware as well, and seems much more likely to stick around in the current environment.
- rektide 4y agoIt's notable that compute modules have had pretty ok availability for a bit. The primary chip seems to be reasonably available. I feel like there's a pretty untold story here, that while yes chips are hard to get & there is limited capacity to a degree, there's also been an enormous amount of buy-outs across the industry & a lot less competition. Not a lot of folk are signing up to try & compete with an rpi like product or main chip, and I think the market's disinterest in serving the low end, in striving to keep going upmarket, is a macro trend that can't last forever.