I'm more waiting to see if a safe method is likely to become available. The people working on Atmosphere were the first to share these issues and it makes sense that they would be working on some kind of solution.
Unfortunately this project is the type that's not going to communicate openly about exactly what they're working on. If you go into the discord, they'll say things like "I can't believe users are doing xyz, they are definitely going to be banned." That kind of attitude wouldn't make sense unless they intended to make a solution that was safe. But there is really no telling with Atmosphere. I don't even think most of us know what kind of solution Atmosphere is going to be outside of "a more comprehensive CFW."
Depending upon what you mean by safe then... I will elaborate anyway.
Can I imagine seeing a fairly workable method wherein all identifying marks for your Switch are scrubbed from a custom firmware type setup (obviously denying you official online in the process unless you inject a working set of keys/serials/whatever the term ends up being and then being banned accordingly)? Absolutely. I expect that in the short term.
Can I imagine seeing a fairly workable method wherein all the checks Nintendo stuffs into the firmware to see if someone is running a modded setup, and all the necessary responses are emulated properly, thus allowing clear online access to Nintendo's stuff? Actually yes I can -- there are setups where it is untenable but as they have to still operate a low power embedded device for consumers and still run high performance games at the same time their hands will eventually be tied and then the infinite time, skills and desire of the hacking set will overcome things. It has happened before, nothing to indicate anything is particularly more troubling here.
Though actually while I said time was infinite there are only seemingly about 50 hackers of suitable skills (maybe 5 times that generally au fait that we have seen in past devices likely to resurface for this) to properly sit down and work this all out. As there are far more interesting things to do first (it is only Nintendo online after all) then at the same time I have not seen (or had any real hint) any of those sit down and pull things apart so as to achieve such a thing.
I should also say that in computer security, and security in general, there is a phrase
"the defenders have to be right every time, the attackers only have to be right once".
Here though the hackers are the defenders -- Nintendo only has to find one little tell (plus all the people sharing hacked lobbies and whatever on public forums) or sneak one little check past the people hacking the firmwares (and we have seen people load far lesser systems with hundreds of checks, often quite hard to spot).
Give me a real budget and a proper team and such slips will not happen but do it in the kind of setups I see hackers typically employ and it becomes a more viable prospect -- to said would be hackers that might be reading this I would expect full diffs, traces, heuristics, semi automated deobfuscation (optimisation techniques here being good), decompilation (obfuscated code tending not to decompile as well) and such to be employed here as well as all the cool things like packet order detection and such like we see on web security (basically becoming a shadow security team). We shall leave discussion of more blackhat methods for another day.
To that end I can see ban waves a la the xbox 360 wherein some hacker makes a slip once or so a year (or Nintendo employs someone with a similar skillset to find a creative detection method) and everybody gets banned before it starts over again.
The comments you see are more likely said devs that might have seen a couple of the checks (but not gone looking for all of them) when otherwise poking things, seeing how a given method will trip the alarms and thus predicting the results. Or if you prefer "I may not know where every trap is but I know where some are and there is definitely one that way".