I'm not sure I've ever liked the term "killer app", because I don't think it's particularly useful in describing real user thinking and behaviour. There was a very long journey from VisiCalc and Wordstar to the modern-day ubiquity of office computing. Different user groups have complex, diverse and overlapping sets of needs and wants that can rarely be distilled into a single application. I'm more inclined to think in terms of Bezos's one-way doors - changes in user behaviour that are sufficiently compelling to be largely irreversible.
I agree that progress has been slow in the consumer space and meaningful long-term adoption of VR has been confined to a few niches; that isn't necessarily an indictment of the long-term prospects for VR, because desktop computers spent much longer in that stage than most people remember.
In enterprise, I think things are more advanced and some user groups have decisively gone through the one-way door for some applications. I think the best example is architecture. If you've done a couple of client presentations in VR, you just aren't going back to showing renders on a flat screen, because immersing the client in a physical space is that powerful. It's not just a sales tool, but a communications tool - clients can understand and respond to the environment intuitively and give much better feedback as a result.
Industrial and clinical training is less clearly one-way, but I think we're very close in a lot of areas. AR is still less developed than VR, but I do think we're on the cusp of something significant - a sufficiently comfortable standalone AR headset with sufficiently high-quality passthrough can deliver training experiences that can't practically be replicated through other means.
I think one of the most interesting areas of development is in psychiatry. It's still early days, but we're starting to see real, meaningful benefits in RCTs for VR-based therapy of disorders like phobia and PTSD. Some of the most compelling results have been in the very sickest patients - people with psychosis, who often find it impossible to engage with conventional psychotherapy.
https://www.psy.ox.ac.uk/research/oxford-cognitive-approache...
I don't think it's remotely likely that VR will ever replace flat screens, but I do think that VR is slowly growing into a niche but durable HCI platform. Tablets are a reasonable analogy - a lot of people see them as a failure, but they still sell in serious volume and they're often a much better form-factor for specific applications than either a phone or a laptop, especially in industry. Tablets didn't change the world, but nor are they likely to go away.