You might be better served learning how bitcoin actually works before trying to "fix" it.
Your ideas fall into either (already planned), (impossible without a hard fork), or (dubious value)
For example there are huge security implications for a node that doesn't have entire blockchain, especially a new node. If you don't want to be a full node then be a lite node and rely on a TRUSTED third party. Using only a portion of the blockchain is the worst of both worlds. You are relying on the honesty of 3rd parties but the third parties you are connecting to are unknown and unverified. Bitcoin treats all incoming data as "suspect" until proven valid. To do that requires the ability to independently validate and you can't do that with a fragment of the blockchain. If you aren't going to validate incoming data then why even have part of the blockchain. Keep 0 GB worth and be a lite node.
There are a lot of flawed assumptions. Transaction fees can never be a % of the sale amount. The network doesn't know the intended sale/tx amount. Fees shouldn't be based on tx value. They should (and are) based on tx size. The critical resource is size thus pricing should be based on size. Fees don't go to node who confirm. It would be nearly impossible to determine which node relayed a tx first. Fees goes to miners as they are doing the significant work of created a block not just tx validation. Coins aren't send to the receiver. There are broadcast to the entire network. There is no way to know if/when the receiving user/client/wallet is aware of the transaction. There is already a mechanism planned for reducing block size called pruning. Miners can reject any tx for any reason (including insufficient fees). If you want to not relay "no fee" tx patching the client to do that is trivial, however remember bitcoin is a mesh network. Transactions will simply route around you.
If a transaction is 'confirmed' x number of times, is there still a reason to keep the original record past a certain amount of future transactions count (or n many months)?
Yes. If you don't understand why then how can you propose fixes? It would be like a mechanic not understanding how a car works but proposing methods to fix problems which don't exist.
I woke up today and had all these thoughts about bitcoin and felt the need to express them. Not sure how much longer I can be a free node.
Then don't.

Still regardless of if you remain a node (no such thing as "free") it might be useful to learn how things work before deciding they need to change. If you aren't a full node then you need to rely on a third party. It is that simple. Most businesses likely will choose to be a full node. Having direct peer access to the network and full verifiable record of all transactions and blocks is worth more than the trivial cost to run a node. If it isn't worth that then you should consider being a light node. There is essentially no cost with the caveat that you no longer are a peer.