No matter how adopted lightning network is, it will not be good for large amount of Bitcoin.
People won't be buying cars or houses through LN.
I don't get that. Why wouldn't it be good for large amounts? Let's say that I have 1000 BTC in an address. Why would it be bad to open a channel depositing 100 BTC and use it for buying cars, houses etc? It's designed for micro-payments, but I'll avoid the mining fees whether I deposit 0.01 or 100 BTC. Someone may want to send and receive huge amounts everyday. He'll find it cheaper if he uses the LN to avoid the mining fees.
There is a reason Lightning Network is referred to as second layer instead of stand-alone network. It relies on "first layer" that is the on-chain transactions
Correct, but I observe a paradox. The first layer requires transactions to operate securely, at least overtime, due to the reward's halvings. The second layer tries to reduce the transactions of the first by having a personal ledger for each user.
So while the second layer makes transactions more private and cheaper, it
digs the pit of the first one.
it will create more on-chain transactions as it is used more since the coins in a LN channel come from on-chain transactions and those channels need to be settled on chain at some point.
The problem isn't that the first layer becomes useless; obviously, it is completely required for the LN. The problem is that it'll decrease the on-chain fees.
Are you running a LN node yourself?
No, but I'm planning to work on it. It seems very interesting.
the rewards will get smaller, but even if the rewards are small in the future, a bitcoin will be much more valuable than today (if everything goes as "planned").
This is a very false thought. The whole system shouldn't rely on the belief that it'll have a greater value in the future. It's important to know what will happen if the value remained the same.