kvstorefool
If you want to check its track record: http://status.mosso.com/
They'll notice when your application is faster and more robust. You don't use a key-value store to impress your users.
Robustness can also lead to easier development, more features, better service, whatever.
Also, say you end up with memcached, or redis, why not cache session data in it so you don't need to query the database or the fs, etc..
Usually there are a lot of things to put into a cache. And memory is always faster than disk.
But that's all assuming you really need it. I wouldn't throw anything new at an app when it's not necessary. You need to know that yourself.
"Using the right tool for the job!" --someone smart
No one claims that key-value stores will entirely replace relational databases. But there are use cases for it where it doesn't make sense to use a relational store, cause no relations are required. E.g., I've personally created a property table in MySQL before. But a key-value store would have been way more suitable for it.
And not just more suitable, but also faster. But don't worry, *SQL excels on other fronts.
Also, Redis is not CouchDB, not even close.
CouchDB is not a key-value store but a documented oriented database. Suitable in case it's hard to normalize your data (or you're fucking tired of doing it again and again). It's also extemely fast on read and easy to integrate (HTTP).
I suggest further reading. Less complaining. Maybe attend one of those conferences. Don't waste your time on Hacker News. =)