Are there any legitimate concerns about tandem breakers?

They are not unsafe.

BUT

If you pick up the tandem breaker and a full size breaker and look at the area that plugs into the bus of the panel and if you take a good look at the panel itself, you will see that there is a lot less area that makes contact between the tandem and the panel bus. In fact about half the area. If the circuit is a heavily used circuit it could be a place that overheats and eventually destroys the breaker and usually damages the bus.

For this reason most experienced electricians like to use full size breakers instead of tandems when they can since they are simply a better breaker. Personally I don't mind using tandems for 15 and 20 amp circuits, but never like to use 240V 30, 40, 50, and 70 amp tandem breakers. Also I try and set up a panel so the heavier loads are not back to back. Usually I try and install them on either the right side or left side of the panel with less loads such as 15 and 20 amp breakers on the opposing side for better distribution and less heat.

I hope this sheds a little light on you question and good luck.

Edit: THIS. Over-full panels invite problems!

Go look at this situation and the poster's proposed solution. This person wants to add a subpanel. What they really want is a 125A subpanel, which should be easy. But they're willing to settle for a 50A subpanel since they feel forced into using a double-stuff breaker, and they're very comfortable using a 30/50 quad; they're mentally married to that solution. But look closer. Look what breaker will be opposite the 30/50. Yeah, a 40/40. So you'll have 160A of load sharing a single bus stab.

And nobody thought anything of this. See how easy that mistake is to make? Notice how a career electrician also missed it - heck, everybody missed it!

A problem like this is extremely rare in a normal panel, because 80A+ breakers are "black swans" - they exist, but you sure know when you're working with one, and so you tend to watch out... and you rarely have two. And they don't sneak up on you.

So that's essentially a UX argument, but that's a real thing in the real world. Armwave all you like about how the world would be if everyone was careful... they aren't.

That's a UX case, but UX cases are why we handle-tie MWBCs, pigtail neutral, use white for always-hot on switch loops, etc. etc. A huge portion of total NEC code is dedicated to UX factors like that.

They're awesome in 2003 or in a FEMA trailer

Because then, you do not need AFCI or GFCI on anything.

But if you provisioned your panel counting on 2 circuits per space, and you then are required to AFCI or GFCI some of those circuits, you'll be gobbling up spaces at twice your expected rate. AFCI/GFCI don't come in double-stuff form factors.

Other emerging technologies won't fit either

Nobody knows what's coming, but I can think of two right now.

• Leviton has a new panel in the marketplace that has integrated power monitoring a-la the Sense home energy monitor. There is a hub built into the panel and smart breakers in each slot. If that takes off, all the panel makers will do it. That kind of thing is usually not offered in double-stuff.

• As we get more into alternative power that is non-dispatchable, we will want to get more into the "smart grid" tech where the power company can suspend your water heater temporarily instead of spinning up a fuel generator to power it. That will be most easily implemented as breakers, and again, not gonna happen in double-stuff.

That last one is already implemented (the brawn part, anyway) in Eaton BRRP or CLRP remote control breakers. All it needs is the electronics to talk to the grid and command the breaker on/off. But one brain could control any number of BRRP/CLRP breakers.

You have more trouble with stab limits

Most panels have "stab limits" limiting all the breakers sharing a bus stab to X amps. Double-stuff breakers cram 4 poles on a stab instead of 2, so they are much more likely to run into stab limit problems. Recently I saw a 50/30 opposite a 30/30, so 140A on those stabs. Stab limits were 125A.