One day, there was a flood in Thailand, and the quality of Western Digital never recovered. The consumer learned that "There are only two kinds of drives -- those that have failed and those that haven't failed yet." So we either spend time trying to remember all the things we wanted to back up with a collection of expensive low-reliability external drives, or employ RAID -- spend more money on purchasing redundant low-reliability drives -- the only reliability that was available since the flood.
And when I say "low reliability," I'm saying that relative to Western Digital's 80 GB IDE drive, or earlier generation that didn't fail unless you hit them with a hammer. Backing up was a simple as a similarly dependable hard drive in an external enclosure, which took some money and maintenance, but it wasn't awful. This is my standard for reliability, for quality, and until we reach that state again, the consumer is being played for a sucker.
Now, instead of spindles and heads wearing out (faster than ever before), there's the threat of the transistors in the SSDs wearing out. There are sufficient reviews by consumers to make it clear that SSDs are a gamble, so it's up to the individual to spend as much money as he has on an SSD that seems to get the best reviews for reliability vs. price. There are only two types of consumers -- those whose SSDs have suddenly failed, and those whose SSDs haven't suddenly failed yet.
What would be good is if the next memory-technology will not have a limitation on lifetime. That would be worthy of being called "an improvement in reliability."