Generally, I'm fully on board with the "eliminate everything but the point" philosophy, but transitions can be really useful for displaying that the same element is present in multiple slides. I've had a few cases where I've shown a "before" diagram/code sample/whatever, then on the next slide shown the same diagram again, alongside an "after" version. But the layout I use to put two diagrams on the screen is going to look different to the layout I use for just one, which means the original diagram is going to jump around suddenly.
A ~100ms transition where the first diagram moves from its place on the first slide to its place on the second slide ensures that a person looking at the slide understands very intuitively which of the two diagrams is the original, and which one has been added. It's not perfect (e.g. you'll miss it if you're not looking at the slides at the time), but for diagrams or code samples you generally want the audience to be focussing on the slides, so it typically works well. And in 90% of cases, even if you do miss it, it'll be obvious after a couple of moments' thought what's going on, but the transition saves you those couple of moments.
I could just show both items on the first slide, but I find it's often pedagogically useful to explore the initial state by itself, rather than jumping straight in with the comparison. That way you can motivate the comparison more clearly by identifying the issues with the initial state (be that a code sample, a diagram, whatever), before moving on to the comparison with a potential solution.
If it weren't for this one use-case, I'd probably also switch to PDFs, because I've been bitten by presentational issues before that would have been a lot easier to solve if the presentation had just been available as a PDF.