Just FYI:
1. Just because someone else is running stable settings doesn't mean you will get stable results even vaguely close. CPUs are "printed" in large wafers, but not every CPU on the wafer is identical: The ones in the centre tend to be more defect free. Defects are not always fatal. The CPUs can be recovered either by disabling the defective core, or underclocking them. So your 13700k wasn't good enough to be a 13900k, but was better than a 13700non-k, or a 13500k/non-k. The thing is, there's still a wide area between being just good enough, and being too good. It seems like your CPU was not at the "too good" end, where as your friends who recommended the 150mv undervolt were.
2. Overclocking is easier in some ways than undervolting. If you overclock, you go from stable at X, to "Will last an hour" at X + 1 to "Might not even finish booting" at X + 2. So most people find when they first become unstable, and back off a bit. Undervolting is much more difficult. It can be very hard to tell when an undervolted CPU is no longer stable, as it can go a long time - even under stress tests, and not crash.
I had a laptop which I undervolted 180mv, and it went weeks without any issues. Started running virtual machines on it (Very load heavy and uses some CPU functions not normally used in a standard PC) and it was randomly crashing. I went nuts trying to figure it out. In the end, to get virtualisation working stably, I had to back off to only 90mv undervolting.
Here's where it gets complex. In theory, it should be no different. Overclocking and undervolting are basically two different points on the same 2D curve, but the curve isn't linear, it's logarithmic/geometric. An example of this is that I got my 8086k to 4.9Ghz all cores on stock voltage. To get to 5.0, I only need 15mv. To get to 5.1, I had to push another 80mv. to get to 5.2, I had to push it to the maximum 1.45v - a full 200mv over stock, and I was throttling and getting heat issues even with a pretty big water cooling system. When you turn that upside down, you can see why if factory was 5.2Ghz, then I could drop a whole 120mv before I ran into any issues, but then at 130, it'll still be stable until all cores were at 5.2Ghz - possibly for an extended period - or if AVX or Virtualisation extensions are used. Hope that makes sense. It might be easier to draw that curve to see why overclocking is easy to test for stability but undervolting is harder.
3. What was stable when you first overclocked the PC, might not be stable a few months later. Room temperature change, dust, even electromigration can change how much a CPU can be overclocked:
https://en.wikipedia.org/wiki/Electromigration You know how a 12v battery can arc across maybe 2mm, but a 40,000v taser can arc across 40 or 50mm? Shrink that down to nanometers and milivolts. You might have 10nm of insulation between two semi-conductors - enough to insulate 1.30v. Over time, electromigration wears down that insulation to 9nm, and it can only insulate against 1.29v, meaning your voltage headroom is 10mv less.
The TL
R version of this post is essentially "If you have unexplained problems, first thing to do is return all settings to stock and see if ti still happens". It's faulty reasoning to think "But nothing else crashes" "But I've been running these settings for years" or "But my friend with identical settings has no problems".
Well, that was an unsolicited novel. Hope it was useful though. I have a degree in computing science, including hardware fundamentals such as processor design, and 25 years of overclocking (30+ if you including changing clock crystals in the 286!)
