About 18 months ago hashrate was ~140 EH/s. Today it is nearly topping 300 EH/s.

Was the bitcoin mining network less secure then?

It's like with a bank vault. If the threat is what cutting torches can do against the vault walls that are made of steel and cement, over a long four day holiday weekend, and it is deemed that 14-inch thick vault walls are more than sufficient for that threat vector -- would you say that vault is less secure than a vault with 20-inch thick walls? Of course not. So today's 300 EH/s doesn't necessarily mean bitcoin's mining network is more secure than when it was 140 EH/s.

The amount of hashrate that exists is essentially the equilibrium between price of bitcoin X issuance rate less the sum of the cost of capital plus the cost of electricity to run that hardware. So a doubling in the price means miner's revenues also double. Then halving occurs and miners are back to the same level they were at before the price doubled. In other words, were are a very long way from where fees will reflect a significant (nonetheless majority) amount of miner's revenues.

Plus there are methods here and more coming to permit transactions to require even less data.

If I am still paying $7 per month to my bank for an account, why wouldn't I be willing to pay a one-time fee of $84 (equvalent to 12 months of bank fees) for my own LN channel, for instance.

But was are sooo far from that yet. The jpegs will lose their novelty, and few will be willing to pay more than a low single digit rate (sats per vByte).

But bitcoin is, long term, going to be a settlement layer only. That's not really a debate topic. So what methods are affordable for the next hundred million (or the next billion) future users to self-custody? Well, just like the cash in my back pocket warrants a different security practice than my main stash for savings, I don't doubt that custodial, federated custody, or other solutions that are not what is recognized as full self-custody will continue to be used for at least a portion of our activity.