Mars isn't important right now. We won't see a human manned Mars mission until the 2040's at minimum because all the details need to be worked out first? It will just be a flyby, like the proposed Venus missions, too. We won't see a human landing on Mars until decades after that. Sure, NASA has vaporware, but when NASA gets going it really gets going, because in the world of man-rated launches, slow and steady wins the race (as SLS proves).
Not sure why you're getting hung up on that though, as NASA is pretty well and truly funding development of LEU nuclear engines for the job of going to Mars, which
you would know if you read the news. If they can take a decade to build a Moon rocket, they can take two decades to build a Mars rocket. Their track record is certainly better than SpaceX's in this regard. NASA works slowly but when it starts working it works out well. It just takes them a while to make up their minds about things.
Conversely, Starship has been in development in some manner since 2007, and was announced in 2005, so it's quite old as far as modern rockets go. It has yet to launch for a shakedown orbit, but has produced plenty of amusing failures on the launchpad and in "testing", which would have killed SLS if it had occurred.
Conversely, SLS has been in development since 2011, as an offshoot of the National Launch System and STS, and uses well-understood and well designed 40-year old technology. Its payload has returned from the Moon to the Earth. Certainly there's reason to question the validity of SpaceX's approach as NASA's rather rapid development has been pretty much on schedule while SpaceX has been foot-dragging.
One thinks if Gwynne Shotwell had gone for a vertically integrated rocket factory that simply mass produces boosters, much like the USSR did, she might be able to have beaten the SLS to the punch. Reusable rockets aren't very useful but they do make USAF contractors happy, although DOD and NASA are very different beasts in their space launch requirements, as anyone who has the briefest look at the history of US space programs knows.
SpaceX is simply built to provide cheap satellite launches for USAF foremost. This translates to being absolutely garbage at delivering reliable man-rated launchers for NASA. Conversely, a reliable man-rated launcher for NASA is a horrendously expensive satellite launcher. The ESA ran into similar issues with its Europa launcher which is why the non-man rated Ariane was selected instead, it was cheaper by a significant amount. No one cares if a satellite explodes on launch but if a person explodes on launch it sort of makes history.
It's sort of amazing that NASA selected the Starship as a basis for its moon landing rocket, but I guess they've gone nuts, or Blue Origin's design was actually somehow even worse.
At this rate it seems like NASA might have to redo the whole lander competition so it can fit into a SLS, probably using Dynetics' design. That will certainly push back any potential Mars mission given how important SLS is to the Mars program. Alternatively, NASA might simply not return to the Moon in the 2020's due to SpaceX's inability to produce a decent orbital rocket and further push back the Mars mission as a result, because returning to the Moon is sort of necessary for Mars in the first place, because the Moon is a good launching platform for a Mars mission (out of Earth's gravity) and close to some cool rocks.
OTOH as it turns out direct ascent has always been bad? Truly, the mission profile rejected for Apollo as being too complex and risky is still too complex and risky, to the surprise of relatively few people. OTOH I guess having semi-functional hardware really helped Shotwell's case
even if she's yet to actually launch it lol. That still leaves time for Starship to flub its launch (not likely but possible), for SpaceX to bungle orbital refueling (equally unlikely, but this is SpaceX we're talking about, second only to Boeing), and for the Starship to fail on reentry or recovery (which has happened with some other rockets) in some manner that puts it in the shop for a while getting a plaster cast on its robot legs.
Remember when people thought SLS would be delayed to 2023? Wild times.
It's unfortunate that we're only getting a sustainable Apollo program now, but Apollo was a one-shot with no real practical support system built up, because it was pure vanity. Sustainability would have missed the deadline of 1969, which is why it was scrapped at the end, and no APA programs went through. Would have been better to take the time to build up a robust space industry and land on the Moon in 1975 or something in a souped up Gemini pod, but hindsight is 20/20 they say.