Interesting and educational post as always ian (and one of the joys of yACF).
Some points I'd like further clarification on if you don't mind and, for disclosure, I'm an industrial chemist and I can interpret data, see trends and understand (basic) bio-chem. Perhaps more relevantly, one of my best friends is a geneticist and I owe most of my knowledge of PCR and virus evolution to him (the rest I learned via Google 
). We've had lots of discussions about Covid. Several* of these discussions have involved alcohol.
"Fact is, SARS-CoV-2 is a very boring virus. It does nothing special."
Yes, but isn't it different enough to infect (and kill) a significant number of a naive population as it would seem to have done ?
That's the thing, we were mostly immunologically naive (possible there's some residual immunity derived from exposure to other coronaviruses, which may explain why some people got it worse than others), so there was little to stop it. Which is likely the standard scenario in a zoonotic spillover. It's basic evolution, throw any organism into a suitable new environment without competition or predation and it'll quickly fill it.
"Equally, I'm not sure the fatalistic messaging that OMG OMICRON IS GOING TO GET YOU! is doomed to anything other than failure because it's not reflective of most people experience of having Covid. We have created a dissonance."
Aren't most if not all of us going to get Covid at some point and the "fatalistic" bit is slightly overdone ? I'm expecting to be infected by Omicrom at some point despite my best intentions. In some ways I'd regard it as a benefit because I assume/hope it would give me a "top up" to my 2xAZ, 1xflu and 1xModerna. Just so long as I don't get very ill and need a hospital bed or die or infect anyone else.
I think there's been some deliberation in the campaigns to drive the message that's it is
you – the individual – who will suffer. That makes some sense, people are selfish, so public health messaging based around people's innate selfishness isn't a bad bet. Now we're in a phase of the pandemic where it's clear that to most people, covid is a minor infection, and that's people's experience, there's a dissonance between what they're being told and what they are seeing. Of course, if you've been vaccinated – even without the booster – your odds of becoming seriously ill with covid are tiny. There's no easy answer to this, of course, other than I think the message is no longer resonant.
"As we can now do unprecedented levels of genomic sequencing (and I'm starting to question the time and money we are now dedicating to this)"
It's the last bit of this quotation that I wonder about. Surely we need to know what variants are out there so that we can see what's happening in the population and we can tailor treatments accordingly ? I'd be interested to know if, by observing the direction of mutations we can predict where the virus might be heading. I suspect not because, as I understand it, mutation is random and sometimes it works for the virus, most times not.
I'm not down on sequencing per se, this is a fantastic advance and – for a start – massively accelerated vaccine development. It's the volume and resources involved in what, in some respects, has become little more than sophisticated stamp collecting. People wiser than me have commented that they could achieve effective genomic surveillance and phylogeny with far fewer sequences. It's also created an obsession with variants. It doesn't really matter what the genomic sequence of a virus is – or how we label any particular collection of mutations – it is the functional consequence of what is a predictably random process. We threw up a wall of immunity and created a huge selection pressure of any variant that isn't affected by that immunity. Again, this is expected – facing massive levels of vaccination and immunity through exposure, either there would be no variant of the virus that could continue to survive (in which case we'd see it disappear) or there are variants that can survive and suddenly have the advantage and can spread fast (which is what happened with omicron).
It's difficult to predict any evolution items since a beneficial mutation is driven by fitness. It has to produce something that benefits the organism in its current environment. Omicron has an advantage because it can spread in individuals with some immunity whereas other variants can't, but omicron might struggle in other environments. Evolution isn't a process of improvement, SARS-CoV-2 isn't getting better. Evolution is a process of survival bias.
"Then we were lucky enough for unprecedented vaccine success (and I don't mean to denigrate the efforts of those involved, but I'm pretty sure that no one would have speculated we'd have debuted not one but several vaccines with >75% efficiency in little over a year)."
My biggest question in your post. Isn't it the case that if this virus had appeared 10, perhaps even 5 years ago, we wouldn't have had the technology to develop vaccines so quickly ? I'm particularly thinking of the mRNA vaccines but even the zoonotic types against Covid 19 ? Again isn't this where our (recently acquired) ability to genetically sequence the initial virus has helped the development of the vaccines and will help modification of future mRNA vaccines ? I'm not sure (from my position of not quite complete ignorance) where AZ go from here unless they can find a different virus to avoid our immune systems and deliver the vaccine.
*A definition of "several" that really means "most".
Sequencing is a lot faster and cheaper, and a ~30 kilobase RNA genome is a straightforward and quick task with modern platforms, so the SARS-CoV-2 genome was sequenced almost immediately by the Chinese. That said, the first full viral sequence was completed in 1976 (admittedly a Herculean task for a virus about a sixth the size of SARS-CoV-2). The 48 kilobase Lamba phage was sequenced in 1981 (but took three years). A significant number of human viruses were fully sequenced in the early to mid 2010s. Sequencing SARS-CoV-2 wouldn't have proved an insurmountable challenge a few years back, though it is one of the largest RNA viruses. Having quick and fast sequencing technology made everything better, of course.
Sequencing works using two main methods – short reads (usually under 500 bases), in which case you chop up DNA into lots of small overlapping sections and sequence those and the reassemble a final larger contiguous sequence and long reads which can process long segments (up to 30 kilobases, though it can be pushed a lot higher) – so fewer overlaps are necessary for a long contiguous sequence or an entire genome. From the mid-2000s, massive parallelization has vastly accelerated sequencing, much of this being technological fallout from the Human Genome Project.
Certainly the ability to go from sequence to vaccine development in literally a few days was massive. You could literally download the sequence in the morning and be working on a vaccine in the afternoon. That was huge (and fortuitous), along with the fact that both nucleic acid-based and viral-vector vaccine development platforms were already established. What, of course, was more fortuitous was that they worked with high efficiency. It's often overlooked, but it's wasn't just a success of developing a vaccine, but the massive challenge of manufacturing and distributing that vaccine.
Whether the success of these platforms is replicated for other diseases is an interesting question that a lot of people are asking (bear in mind that we have spent decades trying and failing to find a vaccine for HIV). I don't think anyone really expected the mRNA vaccines to be so effective (though the principle was sound), in many ways it is an amazingly simple idea of simply jabbing someone with mRNA and having their muscle tissue produce enough protein from that mRNA to generate a good immune response. The viral-vectored platforms were a known safer bet since they were developed essentially as a trojan horse. Virally vectored vaccines, of course, generate an immune response to the delivery vehicle (for AZ, a modified chimp adenovirus), so there are diminishing returns, but the delivery virus can be modified to evade some of this, plus there are other viral delivery platforms).
The downside of both of these is that they are very focused on the single spike protein – other more old-school methods may generate a broader response that makes it more difficult for immune-evading variants to be developed. There's plenty of work ongoing to create vaccines using these technologies that generate multiple epitopes (variants of the spike protein or other components of the SARS-CoV-2 virus). These would, in theory, dramatically limit the probability of immune-evading variants, since they'd require a series of mutations that are improbably and if they did simultaneously occur, they'd deplete the overall fitness of the virus so much that it would likely not have much of a future.
