It's not stuck, you can still spend it.
How? if the fee becomes consistently higher than the amount held in the address.
You might be able to work out a deal with a mining pool in which you provide them with a transaction with inputs and no outputs and the pool crediting some account of yours a portion of the value of what was "donated" to the pool, although this would only make sense if the pool owed you money as a result of your normal business activity.
This is a very good reason to advocate for a larger maximum block size.
The money is actually not lost... as suggested by HeRetiK above, you could just send a bigger amount of BTC above the resultant fee and mix both coinbases to join a transaction that includes the money that's stuck on this wallet which only got this only one address. Of course, you have to consider if it's worth it as you would lose money in fees doing this in each address that's stuck somewhere you had a wallet with enough funds to transact it out.
It's not stuck, you can still spend it.
How? if the fee becomes consistently higher than the amount held in the address.
Try consolidating these few Satoshis with some of your other coins by sending them with an ultra-low fee to another wallet you own. It may take some days, but they will be arrive sooner or later. If your transaction gets dropped from the mempool after 1-2 weeks, just try again. It'll get through eventually.
In LN, you need to create a funding transaction which sends the money you want to put into the channel into a special 2-of-2 multisig address. From there, you can do microtransactions with the money you deposited.
I got the idea, but what happens when you can't afford the first on-chain transaction? If the idea is "banking the unbanked" as Andreas Antonopoulos goes around saying, how can some poor person in good knows where, afford that first on-chain transaction? Let's say average on-chain transaction reaches $50-100 in the future and you want to bank the unbanked in africa or who knows where else. Im not sure if they will have that $50-100 entry fee (+ the amount you want to use for microtransactions) lying around.
It's a valid concern, but I honestly doubt that on-chain transactions will reach fees of that magnitude once lightning network gets deployed. Even if that were the case, people would most likely simply use alternative crypto-currencies until they reach capacity as well.
Fees will go up on-chain with LN, as far as I can predict. Since LN is just packaged on-chain transactions, if a ton of people use LN, the overall on-chain fee goes up, which will be annoying for someone that
1) Doesn't want a bigger blocksize
2) Isn't really interested in microtransacting over BTC
Since when I need to make an on-chain transaction, the fee will be higher due LN usage, which is why I would rather see LN over Litecoin, and leave BTC as digital gold only.
