The Lightning Network is also very interesting tech.  I highly recommend the whitepaper to anyone with a undergrad background in CS or math: https://lightning.network/lightning-network-paper.pdf

IMHO, any solution that requires trust in a centralized third party is absolutely inappropriate for a cryptocurrency.  The whole purpose of cryptocurrency is to eliminate the need for trusting centralized third parties.  A cryptocurrency that depends on centralized trust is the worst of both worlds: it has both the necessary inefficiencies needed for a trustless system and the failure modes that arise from centralized trust.

This is absolute nonsense.  See the 16th Amendment and 26 U.S. Code § 1.  If you're a US citizen who has realized capital gains on cryptocurrency, you are required to report your gains as income and pay taxes on them.  Exchanges which accept deposits from and withdrawals to US bank accounts can be expected to comply with IRS demands for information on their customers.  If you evade paying taxes on substantial cryptocurrency gains, there is a good chance that the IRS will find out eventually and you will be subject to civil and/or criminal penalties.

At market equilibrium, the expected benefit (revenue) from Bitcoin and other PoW mining would be roughly equal to the expected cost (including electric power cost).

A downside to proof-of-stake systems is that it is a "rich get richer" system: Only people who already have a lot of the currency have a decent chance of getting a block reward of more currency.  PoW is more egalitarian: Anyone can buy a mining rig, and they don't need to compete proportional to the currency holdings of whales but only proportional to their investment in mining (hardware + electricity).

Unfortunately the Bitcoin community hasn't been able to agree on a short-term solution to high transaction fees, but hopefully in the long run, the Lightning Network will be able to reduce the number of transactions that need to be on-chain, thereby reducing the fees.

The difficulty adjustment is slow.  It will take weeks for Bitcoin block generation to adjust to a significantly smaller global hash rate.

I'm going to hold both.  Every time I've sold Bitcoin for fiat, I've regretted it later.

Right.  Publishing a transaction to the network of one fork will not automatically publish to the network of the other fork.  However, a replay on the other network doesn't need an 'attacker' in the sense of someone intentionally being malicious.  It possible for a misconfigured node to accidentally publish it to the other network.

If you're seriously suggesting hacking someone's computer to steal their private keys, well, that's an exceptionally bad idea.  It's a felony under the Computer Fraud and Abuse Act (CFAA) and can land you in federal prison if you're subject to US jurisdiction (and most other countries have similar laws). Cryptocurrencies may seem like a wild west, but there are more than enough laws on the book to punish cyber-criminals.  The government is slow-moving, but it is gearing up on cryptocurrencies, and soon enough you'll start to see prosecutions of hackers who steal Bitcoin.

You're going to have to wait a bit until the Bitcoin Gold network is operational.  Right now, there is no method for splitting that has been tested and validated.  As for exchanges, we will see in the coming weeks which, if any, exchanges decide to support Bitcoin Gold.  I personally would not be surprised if Bitcoin Gold is a complete flop with negligible market value and no support on major exchanges.

One possibility is that advances in cryptanalysis or quantum computing eventually break SHA256 or ECDSA secp256k1.  With any luck, this would be gradual process (as with MD5 and SHA1), so that there is time for an orderly hard fork to a different digital signature algorithm or cryptographic hash algorithm.  Then the UTXOs with lost private keys can either be formally destroyed as part of the hard fork or claimed by whoever first cracks them.

And as Saint-loup notes, if Moore's Law continues long enough, even brute force attacks might eventually compromise the cryptographic primitives currently used by Bitcoin.