I get it. I am really excited about the current state of open source FPGA tooling, along with newly inexpensive and capable FPGAs as well as new low cost foundry shuttle services. Also the massive productivity boost LLMs provide. Feels like I have the world's most capable army of software development interns for $20/mo.

Projects like MiSTer are very inspiring. Risc-V as well. Sam Zeloof's garage chip fab work too. And we even have reasonable platforms for developing open source phone stacks like Pinephone - I remember the bad old days of OpenMoko.

I think proprietary chips and boards are about to go the way of proprietary *nix. It'll take a decade or more, and lots of work. But the future's never looked brighter for open systems.

the issue i think is not the design but more the fabs. there are too many designs and not enough sub 10nm fabs

There will always be a premium on latest node fabs. Nothing to be done about that without billions of dollars to invest, which comes with it's own strings. In time sub-10nm fabs will be older and less expensive as newer nodes come online.

I don't need the fastest or lowest power devices though. I'd happily trade some of each for a more flexible future-proof machine. I just need an FPGA big enough to hold a linux-capable core or two, with graphics and audio and networking at an affordable price. Bonus points if it has some extra space for developing new peripherals.

I think it'd be pretty easy to design something to conform to the raspberry pi compute module interface, for example, which would make it a drop-in replacement for lots of useful systems like laptops, NUCs, and other such stuff. Gotta love defacto standard interfaces.