Any half way competent DM knows this is a feature not a bug. So the issues that arise are not failures of the system. I know there’s a certain subset of DMs that don’t think they should be responsible for their games.
Edit: the real failure here is the idea that a DM must ‘work around’ their players strengths. Those strengths are something to work with.
Right, so if you don't think it's a feature, you're an incompetent DM? Sure, mate.
Look, I have no problem running it as is -- I've got three campaigns under my belt that haven't touched expertise and run into the middle teens. I can ignore the problem just fine, but I don't pat myself on the back and assume I have mad DMing skills because I can ignore the problem. Expertise as a mechanic just sucks. It's a purely numerical boost in a system that otherwise works to reduce purely mechanics adds. By itself, it's only annoying -- it rapidly renders challenged trivial. Whatever, that's easy to pivot on and switch what's challenging as it happens. It's unfortunate that it does it in fits and spurts (+2 at a time), but, again, you can work with that. When it's also paired with all the other skill boosting mechanics, though -- it just becomes number porn.
Frankly, understanding expertise and how it impacts the game requires grasping the incentives the system has. Expertise, because it's so much better than just about anything else, creates an incentive to get it as earlier and often as possible because most people run games where skills are binary. You have to do a lot of work as a DM to run a fail-forward style that the players have enough faith in that they aren't stacking expertise on top of bardic inspiration on top of guidance. The incentive is to eliminate any chance of failure, and expertise is the best method of doing so available. When you stack that on top of reliable skill (another bad mechanic in execution), you end up with an incentive for the player to best the game by rendering entire challenge areas obsolete. Can the DM overcome this by introducing new challenge areas and abandoning old positions? Absolutely, but that's creating a workaround for a bad mechanics -- you're replacing skill at knowing when the system breaks and going to something else for a good system. This is the hallmark of a stealthy bad mechanic.
Expertise should reduce the risk of failing at easy or moderate challenges, not make difficult challenges trivial. It should have been advantage, not the class-disconnected double proficiency. This would have achieved the design goal easily, and not allow tier III and IV characters to make the DC system obsolete for a swath of very useful skills (the unbalanced nature of skill usefulness is a different topic, and one I don't think can be solved without a completely different game, which there already are many of, so why bother). Reliable talent can stay the same under this and provide even better protection against flubs at moderate to difficult tasks (it autos easy, which is absolutely great!).
The incentives that expertise causes (especially on top of reliable talent) are exactly what's already been said by the other -mancer (represent!) -- it discourages these things from even being challenges. The DM gets nothing from doing work to add such challenges if they're trivially dealt with at the table, and the player gets scant spotlight time for saying 'yup, auto-did that again.' Again, this is further exasperated by a binary pass/fail skill paradigm (which is pretty common and is what shows up in official adventure material). This pushes players to engage in build choices that remove the chances of failure in mechanical ways while also reducing the impact (and therefore usage) by DMs of those elements. It's a double-edged sword with many confounders, but, ultimately, the mechanic is at fault. Use of expertise on top of reliable talent will absolutely result in fewer of those elements being put into adventures because it becomes boring to just do the mental masturbation of saying 'there's an X, you bypass it, don't roll'. Denying this because you can do this workaround and instead focus on other things doesn't mean it's not an actual issue.