Very much other for me because it needs to be divided into two separate categories: FPGA replacements for the standard TI-99/4A (and possibly related/similar machines), vs something new that happens to kind of be backward compatible.
I don't object to recreations and updated versions of things--the supply is limited and if these things are to be preserved, then the actual hardware which won't last forever needs to be preserved as well.
For the time being, there are a lot of consoles out there, and they're cheap and easy to find. Ten, twenty years from now, who's to say our GROMs won't all have suffered from bit rot or some of the custom chips in these machines may no longer be working or replaceable.
If for that reason alone, I favor FPGA and other sorts of recreations focused on accurately recreating the console.
For the TI, that's going to mean a need to recreate the sidecar and cart connectors, and allow for a lot of configuration of the FPGA's recreation of the system. You're essentially going to want a configuration screen with a lot of knobs. People have had 32k RAM in sidecars or PEB or nanoPEB, but some folks have the option of making the RAM be 16 bit for speed, for example. And 32k or 1MB or no expansion might be desirable options. And what hardware is or isn't present? If it's not part of the FPGA (and it might be very hard for it to be), MBX folks would like to be able to connect that device as well. There's a ZX Spectrum FPGA board about to ship for < $100 in its basic configuration that can pretty much replace any existing Spectrum and for the first time fits into the 48k Speccy shell--I dunno if I'd want that for the TI-99/4A since the Speccy is small, there are lots of empty shells with long dead boards, and there are now new 3rd party shells so no cannibalism is necessary.
That last one is big for me: I don't like raiding working hardware for parts. If the parts are available separately or there's NOS or even just new stock available, that's another thing entirely. But if a part is not made anymore and custom, it feels WRONG to pilfer from one machine for another to me. (That's why I've been on this GROM thing on FB lately..)
And it leads to the other issue: The Super-TI-99. You can make it as tricked out as you like, I suppose, but I consider base compatibility to be the #1 goal. If you can double the CPU speed and give it access to more RAM or GROM than anybody knew what to do with in 1983, go for it. F18A, sure. An "F18A for sound", sure. But all of that has to come after the TI itself is recreated.
And I really feel like the result needs to target a US$200 price point. That might be hard to do with an FPGA that can truly do all the things, but if it costs $300+ and the developer is losing money at that price like with the Turbo Chameleon 64, that's not a good thing. Lots of folks would love that in the Commodore community, but they just can't justify the money, and Commodore's got a much larger community. There's a Commodore UG that meets within bus distance of me every month. Man, for the TI, we get together twice a year in different parts of the country. We don't have the userbase for something to be successful if it winds up being something that not everyone's gonna want (purists are a thing and what they do is respectable even if I can't justify it myself) and those who do want it have a hard time justifying the cost.
Just random thoughts from a wandering albino who needs to read the forum more often.