Personally, if I am doing any kind of technical diving, I want a buddy who is smart enough to figure out what hose I am on. If he is calm, he can help himself to the long hose, if I see him coming, I'll have the long hose in my hand to give to him. If he needs a reg right now and I am on the short hose, he can still take the reg out of my mouth, I can get the long hose and then we can switch him to the long hose.
For the most part it really is not an issue. If you are paying attention to your buddy, you will notice when he encounters a problem with his reg and can adjust accordingly.
Plus if I have an OOG buddy on a techical dive, it's the last time that moron is diving with me. I come from the other wschool of thought that each diver should configure and plan to be fully independent in terms of gas/regulator related failures and should have the means to get themselves back to the surface or deco gas. A buddy is at most a backup to the backup plan and if you have to use one to access gas, you really screwed it up. There is nothing wrong with team diving, it just is not an excuse or substitution for proper redundant configuration and planning.
I do agree with you that if I have a regulator failure on indepenedent doubles, I will lose access to all the gas in one tank. However with proper gas planning and management that still is not an issue as what's left in either tank will get you back to the surface or the first deco gas switch.
I do not agree that failures are more likely. At a minimum reg failures will happen with equal frequency regardless of whether the regs are on two single posts or on a manifold. In fact, I'd argue that the regs on independent doubles are likely to be more reliable as both are used equally during each dive. One is not used more heavily than the other and one is not essentially ignored on every dive and both factors have an effect on reliability. I used to have to swap regs mid season from manifold post to post to even out the use over the season.
A tank o-ring failure or blown burst disc is also a wash as a manifold will not preserve or allow the use of that gas any more effectively than with independent doubles.
So what you have left is the potential failure of an o-ring, seal or valve in the manifold itself and if you start counting o-rings and valves, a manifold will have more than a set of independent doubles every time. So that would be more potential failure points, not less.
Plus, lets remember that when you have a manifold with an open isolator valve, you are required to close the proper post and/or isolator in the event of a failure to prevent the loss of all your gas. In a restriction, you are going to be a lot more stressed and nervous about a freeflowing reg than I am going to be. In fact in some cases, you could be really screwed.
So in the end, I think the pros and cons pretty much even out. You have potentially more gas with a manifold in the case of a reg failure (but no other failure) and with independent doubles you have less potential for failure as well as the assurance that no immediate action is required in the event of a failure to save your gas supply.
So I'll continue to dive independent doubles and hedge my bets on the potential reg failures by diving with well maintained and very reliable regulators.