I was thinking the same thing. You only see posts or hear about issues. No one with a perfectly working drive takes the time and effort to post in a support forum.
a few dead drives is nothing. I bet there is a few dead drives each day with Samsung... probably about 3% of sales. That seems about normal for SSD.
You can't declare anything on such a small sample size.
A few things:
1) Plenty of people with good drives do post in threads. Lots of people post their benchmark scores, ask questions about how to increase performance, discuss performance expectations, etc. There's plenty of discussion in threads where you can infer a drive is functioning from the context of the comments, even if the comment isn't itself "my drive is working perfectly." I'm not saying that most people with good drives post in threads - they don't. But I think the notion that bad experiences are over-represented in comments while good experiences are under-represented is not necessarily true.
2) Under-representation in public threads cuts both ways - if good experiences are under-represented in them, so too are bad experiences. For every one person who has the knowledge and inclination to post about his/her problem, there are lots of folks with problem who don't. Not everyone with a problem has the awareness or even desire to post on forums about their tech problems. Lots of people with problem with deal privately with the retailer or tech support from the manufacturer and never discuss it in public forums.
3) It's always the case that public reports of a problem undercount the actual occurrences of the problem. This is not just true of tech problems but all walks of life - the vast majority of crimes are never reported, the vast majority of diseases or outbreaks are never publicly counted, etc. What's more, not only are problems under-reported, most people may not ever even experience a failure even if their product has a known defect. The vast majority of IBM Deathstar drives lasted their useful life without ever dying from the click of death (I wasn't so lucky). Even in the similar Iomega Zip Drive click of death case in the 1990s, where thousands of complaints were reported, Iomega said that 1 in 200 drives were affected - 0.5% (this according to Wikipedia's writeup of the case). Iomega sold millions of those cartridges. There was a design defect in those drives but still, most of those drives never broke down and even the reported dead drives was a tiny fraction of the total numbers sold. That doesn't mean those cartridges didn't have a design defect that made them vulnerable to read/write head failure.
I'm not saying every report of a problem is evidence of a fundamental flaw but every report of a problem should to be treated seriously. Look at the confirmed defects in the pre-production firmware of the Samsung 840 drives. As far as I can tell, a grand total of six dead drives triggered that investigation - two at Anand's labs, four elsewhere. Six drives out of thousands in the hands of end users. Let's be conservative and say Samsung had sold 1,000 840 drives at that point - 6 dead drives is a mere 0.6%, far below the 3% failure rate Mfusick describes as "normal" for SSD drives. If Samsung had an equally relaxed attitude toward SSD drive failure, it would have never discovered the bad pre-production firmware bug.
And btw, one of the reason I first went with the Vector over the Samsung 840 Pro is because of those reports of dead 840 drives. I read Samsung's assurances that the bug was only in pre-production firmware and did not involve retail products but I wasn't completely convinced (and honestly, I still sort of expect the Samsung drive to die on me). I think anyone who chooses to still pass on Samsung 840 drives because of those drive failures would be pretty reasonable.
