Has Prusa given up on the MMU?
The repository for the actual parts has had 4 commits total since the first one in 2018. Two of those were adding printable parts for the filament buffer. One was for changing the identifier printed on the parts. Only the last, an improvement to the blade holder, is even conceivably about improving the MMU's performance in any way, and the general consensus seems to be that if you're relying on the blade you're already doing things horribly wrong.
The last firmware update was almost a year ago, and again, didn't actually address the most common issues.
Despite new threads just about daily, there's no real official documentation on how to actually get good tips and consistant loads/unloads, rather than any authoritative guidance, we're left to look through a half dozen unofficial guides across peoples personal web pages and forums and reddit and hope that something will work.
I know these are just 'community forums', but at what point is the volume of complaints sufficient that Prusa might actually step in and say 'we hear you'? Or are they just hoping we'll all eventually just give up and toss it in the trash and forget all about this?
You can contact the support which points you to their existing documentation which is nice but does not help in case of MMU2S. You read all the forums, prints different modified parts and try to replace the standard parts hoping it will resolve your issues. And one day you just give up, stop using the MMU2S and look for another (and dual extruder) solution as I did. Prusa does not want to build the box for the printer? OK. Prusa does not want to fix the slicer issues? OK. Prusa does not want to fix the MMU2S issues? OK, there are other printers on the market. I will wait till end of the year and if Prusa will not come with something bigger (fixing all MMU2S issues, new big printer or new interesting products), then I will just sell the MK3S and MMU2S units I have.
3.9.0 for the printer contained some MMU bits.
I don't think that they've given up on it, but I think it is a fairly low priority and has been swamped by the SL1, Mini, and large format printer to be.
I think the biggest issue with the MMU is also the hardest for Prusa to solve. It comes down to everyone's configuration being different. Seemingly insignificant differences can make a big difference due to the impact it has on filament movement.
The one thing I think they could fix is to add a LED on the IR sensor from the factory. While my obsessive war on friction helped greatly, being able to clearly see what the IR sensor is doing in real time is what moved it being reliable.
The IR sensor issue is admittedly one of my biggest complaints - My main failure mode seems to be where it loads the filament perfectly, according to the screen the sensor is good, but evidently it decides something isn't right and backs it out again. I'd add an LED myself if my soldering skills wern't complete garbage and if I wasn't so bitter about the fact that something so obviously essential to debugging was left out in favor of a screen on the LCD that's evidently too inaccurate to be useful. (Though I'll probably just admit defeat here sooner or later and at least try to make something work)
But really, it wouldn't be so frustrating if Prusa would just at least address some of the complaints and at very least give some official recommendations. For example your tips in particular have been pretty helpful, but I've also seen people elsewhere saying that higher ID tubing can actually cause more friction issues, or that you shouldn't make any modifications at all because it works best completely stock, with Prusa's radio silence it's hard to know what to think.