Jeff Bozanic was one of the people pushing that argument 4-5 years ago. I agree to an extent in that it is a bit more idiot proof as without the isolator, a diver can't close it at the wrong time, fail to check it pre-dive, etc.
But that is also a bit like saying that firing a gun with a safety is more problematic as the gun has a safety that can be left on, left off, or fail. That's all true, but it must be balanced with the greater overall safety that is attained and any added faults can be addressed with better training and/or maintenance.
Instances of neck o-ring extrusion, burst disc failure, or manifold failure have a very low incidence of occurring - but the total gas loss that results creates a significant risk of a fatality, especially in a two person team diving thirds and near the point of max penetration.
On the other side of the coin, actual failure of the isolator valve itself is equally rare - I have never actually heard of one. At worst, it adds one more potential very low incidence failure to the existing list of three, but it makes those three much more surviveable.
It also adds some potential failures due mis-handling of the isolator valve such as starting a dive with it closed, breathing off the right tank and reading the SPG on the un used left tank. However in that case and in most others, good pre-dive checks and good situational awareness - a truly attentive diver will not go far before they notice they are using a lot less gas than normal and check the isolator to discover it is closed. In that regard, I think the early problems with isolator manifolds were due to existing pre-isolator habits and and training protocols based on older non-isolator techniques.
Personally, I check my valves at least twice in the pre-dive phase, can reach them underwater and am alert to any oddities in gas usage that may suggest a closed isolator, so I would rather have the ability to isolate in the event of one of the unlikely failure modes than to have a more idiot proof manifold that cannot be isolated.
----
In regard to when it was first invented, Gilliam states "late 1960's".