Mine has been on both of my new MBPs. In fact, they've been just fine for the majority of users. If it was as huge of a problem as some here seem to believe, you'd have seen mass recalls and it publicized all over national news.
The reality is that only a very small number of users have been impacted by this.
Wrong. Mass recalls happen when there is a
safety issue, not merely a failure to work properly. Likely nobody will get killed by a MBP keyboard failure, so that's why you're not seeing "mass recalls and it publicized all over national news". Elsewhere in this thread there are reports of large companies seeing failure rates around 14%. That's roughly 1 in 7. Now, if you are an isolated user (not talking to a corporate IT department or similar) and you happen to be amongst the 6 in 7 who aren't seeing any problems, it's easy to say the problem must be user error or the problem doesn't really exist. And 6 out of 7 working is, indeed, the majority of users. Of course, for every person taking their machine in for repair, there are likely others who are having occasional problems and simply ignoring it, overlooking it, or assuming it's their own fault (e.g. that they must not have hit "g" squarely, because one
expects a keyboard to properly transcribe exactly what they type).
Now, by definition, a product that works for 51% of the users is "just fine for the majority of users". Would you call a product that fails 49% of the time "just fine"? Or would you call it alarming? Uh huh. The test of "
works for the majority" is
not a reasonable test. (FWIW, the huge problems with airbags over the past half decade have been in relation to airbags that were working about half the time - and failing about half the time).
You want the failure rate somewhere down around (ideally substantially under) 1%, not at 14%, and especially not on a piece of hardware that is a key part of a user's workflow, isn't instantly interchangeable (sure, you have backups - assuming you have a replacement machine to swap in, loading from backup and getting back in action takes quite some time), takes multiple days offsite to fix, and costs a substantial fraction of the original cost of the machine in order to fix.
If you aren't having trouble with your keyboards, good for you - but don't try to minimize those who
are having trouble (you're working from a sample size of 2 and yet faulting others for using anecdotal data). It sounds roughly equivalent to saying "I don't have cancer, and the majority of people don't, so cancer isn't really a problem."