The problem here is that with two tests that must be achieved to sneak, it makes it a lot harder for the stealthy PC. Now try to handle it with the rogue trying to sneak past multiple potential opponents - UGH! And that's ultimately the trouble with multiple checks being necessary to succeed at a task - it becomes harder to do.
It’s a sensible argument, but in practice I do not think it was really a problem. Let me explain…
The problem would be significant if we took 5e and just split stealth and perception into two skills each, while keeping everything else as is. But in 3.5, there were many other mechanics that interacted favorably with this.
Let’s say, hypothetically, that we consider the case of a character with a dex of 18 (+4), ok? Well, in 5e, a 1st level character proficient in stealth gets another +2, for a total of +6 (1.5x what the stat gave) in stealth, while a 20th level character gets a proficiency bonus of +6, for a total of +10 (2.5x the stat). If they have expertise too, then +16 (4x the stat).
Contrast with 3.5, where a 1st level character could put 4 ranks each into move silently and hide in shadows, for a total of +8 (2x what the stat gave). At 20th level, that character could have 23 ranks, for a total of +27 (>6x better than just the stat).
I’m assuming no stat increases to simplify the example, but even with ASIs, the point ends up being the same, which is that in 3.5 the training counted for way more than the "inherent" ability. Why does this matter though? Well, that should become clear in the next point…
In 5e, all characters have almost identical amount of skills. The "skill monkey" classes get a little bit extra from expertise, but it’s fairly limited. Compare with 3.5 where a rogue would get 8 skill points per level, a bard 6, and many other classes got just 2. This means that for a low-skills class, it was a big opportunity cost to invest in the two detection skills. Maybe you’d invest in just one of them but not both (e.g., the keen-eared fighter and the sharp eyed ranger, adding to the flair of these characters).
Because of all these factors, I would argue that a rogue was not necessarily penalized, since they had available to them plenty of resources (skill points) to invest in training the two aspects of their stealth significantly more than their adversaries could spend training on the two aspects of detection.
Now, if we scale the scenario to multiple opponents, it doesn’t really change much. Fighting multiple opponents would be harder. Going past them stealthily should also be harder (unless… being a lone guard is boring, but there is nothing else to do besides looking around, whereas two guards can have a much better time gossiping, gambling their wages on a dice game, or any other distraction!)
