The 4.3 in my S10 would do the same thing. It would feel like the engine died for a split second and not just a single cylinder misfire. It did this after I had floored it (accelerating, merging, passing...) then went back to a steady cruise. It would do it just once. If I gave it a little throttle, like just enough to accelerate a bit but not downshift, then settled at a steady cruise, it would do it once more. I could stop, shut it off and restart it and it would fix it until the next time I floored it then cruised.
Your 350 is essentially the same in the front and probably has the same timing cover with the sloppy manufacturing tolerances. The CKP sensor doesn't have a very wide tolerance for spacing from the crankshaft. Some covers put the sensor too close, others too far. Most were too close which is why the spacer was introduced. Of course, there's also variances in the CKP sensor manufacturing itself to further exacerbate the problem.
With my issue, all was fine until the crank moved enough radially at high RPM to contact the sensor. My guess is that this threw it out of whack and the PCM kinda resorted to a base map to keep it running normally. Applying throttle causes a change in RPM and the PCM alters the timing, so this is probably where the hiccup occurred. It's like the PCM just cuts all ignition activity then resorts to it's base map which gets it running again. This happens so fast because of the processor speed of the PCM, so it doesn't seem like a single- or multiple-cylinder misfire and more like a quick "key off/key on" action.
If you still have access to that MAC scan tool, have a CKP Relearn performed after you inspect (and space out if necessary) the CKP sensor. You may be able to fix this with no parts cost.