I don't know if you are joking or forgot about SPV mining. It isn't that simple.
If mining in SPV-mode produces valid blocks - it can be treated as holding full-node.
I do not see any problems here.
OK, now I understand that you have an oddball sense of humor.
The "full node" is traditionally defined as a node that:
1) knows the entire blockchain history
2) propagates transactions and blocks after verification
The "pro-pools" legacy meaning included one more function:
3) mines blocks
Bitcoin started in the situation where mined coins paid for all 3 things:
3) electricity spent on PoW
2) network bandwidth
1) block storage
Now that node specialization is progressing I think the first two functions will get unbundled too. I thus don't consider "full node" a very important concept. Perhaps it would be better to split it into 2 or 3 terms:
i) real-time propagating node
ii) UTxO query-service node
iii) archival query-service node
depending on how far we would want to go into unbundling.
The problem with unbundling is that there's nobody paying the fourth cost:
0) research and development
Therefore after mining got unbundled the developers were left to funding themselves by extortion, scams, rip-offs, vulture/venture capital, preying on the naïve and other miscellaneous funding sources.