Is increasing the block size really a solution? Let's imagine that the block size is increased significantly. Even then, each transaction will take around 10 minutes to be validated on the network (and this is an average, it can sometimes take up to 1 or 2 hours), so I don't see how increasing the block size would solve the scalability problem. In fact, on average, the amortised analysis will tell us that the network will be capable of carrying out thousands/millions of transactions per second, but we'll still have to wait for the transaction to be finally validated. A more logical solution would be to reduce the time between blocks (without touching the supply, all we would have to do is divide the reward for blocks by 100 if we want an average time of 6 seconds per block, for example), but this is obviously unthinkable given the decentralisation problem it would pose. But if one day the price of hard disks were to be divided by 100, or by 1000, would that still be a problem? A 100TB hard drive for $50, for example. As we saw above, you'd need to get a network that's scalable enough for 9 billion people, and without doing any calculations that seems very complicated to me. I prefer a second-layer solution like LN, but for the moment there's no miracle solution.