The USB standard (v2.0) calls for minimum 1500 cycles at 200 cycles/hr maximum. Honestly, this is probably comparable to the cheap 79c NEMA 5-15 outlets you get at the hardware store. "Commercial Grade" outlets will probably exceed that by a fair margin, but higher quality USB cable assemblies are also available.
The locking mechanism on micro USB isn't because the full-size connectors wear out too quickly, it's because the connector is too small to have a very reliable friction retention mechanism.
I've never actually worn out a full size USB A or B connector, though I'm sure it's possible. The design is actually pretty darned rugged; the only thing I can see wearing out very quickly at all is perhaps the retention clips, depending on how they're constructed (they could lose their spring).
What concerns me more is that the USB A connector is just a bad connector on several other points. It's not capable of very high current (500mA by the original spec; battery charging spec brings it up to 2A, but many connectors are actually only specified for 1.5A. This is only 10W (a linked article from the linked article indicates that there may be an effort to fix this, but it may just be the old USB + Power standard mentioned below - I can't tell since the link to the USB IF document is 404). There are also no provisions for adjustable voltage, which would be handy in a real low voltage DC power spec. It's 5V (and with pretty wide tolerance) or bust. The connector also suffers from being rectangular but requiring a specific orientation not immediately obvious from the physical shape. There are some guidelines for identifying the proper orientation using overmold artwork and similar, but these are regularly disregarded or ever flat out violated.
There is a "USB + Power" standard, but it never really caught on. It actually fixed all the problems I highlighted above at the expense of adding another part to the USB connector to transfer the high power.
As for isolation between the AC line voltage and low voltage stuff, it should be doable. If the low voltage is generated within the outlet housing itself, that's up to the manufacturer to get right (and, being a fully controlled environment, should be easy enough), and there's no problem in the box. That just leaves the possibility of a contact event on the plugs themselves. The linked article is probably just a concept drawing, but that would be a pretty terrible way to do it (they did at least put the connector on the neutral blade side). Proper recesses could make a contact event quite unlikely. It's such a shame that the US never adopted the popular Japanese practice of having the plug blades partially sheathed in plastic so that the exposed portion is insulated by the time electrical contact is internally made (like the EU Schuko plug).