Everyone's experiences are different of course, but I've seen quite a few broken bolts over the years with direct impingement AR's, besides gas rings, tubes, extractors, and various springs burnt out way too early. A direct impingement is one of the toughest rifle on inner receiver parts I've seen and the most lubrication and cleaning needy. I've also seen AR's take a lot of beating and round count and not have any problems with parts change out being much later, but that doesn't change the fact of all the others that did. Still both an AR-15 and a .308 AR can be a good rifle if you give it what it needs, when it needs it. I prefer other .308/7.62x51mm rifles like the M14, G3, and FAL more and even though I've had various AR-15's over the years, I prefer piston AR's these days. There are choices out there and it's nice to have them.
You might look at these also,
This is from that article by the way,
At the 5,000 round mark, the bolt carriers, upper receivers, and barrels were cleaned. After observation of high speed video showed inconsistent cycling, action springs ($3) were replaced, as were extractor springs ($6.99) and gas rings ($2.19).
The second half of the test started off with several malfunctions with the Brown Bear carbine - at 5,200 and 5,250 rounds, short stroking malfunctions were encountered. High speed video showed that the bolt was barely coming back far enough to pick up the next round, and occasionally not even far enough to eject the spent case. Additional lubrication did not prevent the second malfunction.
A detailed physical examination revealed previously unnoticed carbon buildup in the gas key and gas tube which had almost completely occluded those components. The other firearms were inspected, and none exhibited carbon buildup which was even remotely close to that of the Brown Bear carbine. Cleaning of these components in the field proved difficult to impossible, and it was decided to set them aside in order to examine the phenomenon.
The gas tube and bolt carrier of the Brown Bear rifle were replaced with identical components, after which firing resumed without incident. No malfunctions occurred until 7,500 rounds, when five stuck cases were encountered between 7,500 and 8,200 rounds. From 7,500 rounds on, a number of cases with distended and/or split necks were observed.
The carbine firing Tula had a case stuck in the chamber after 189 rounds which proved exceptionally difficult to clear, even with the use of a steel cleaning rod after the rifle had cooled. Over the next three hundred rounds, 24 malfunctions - stuck cases and failures to fully cycle, or "short stroking" - were encountered. At this time, the Tula carbine was removed from the testing, as the problems were causing significant delays.