Seems to be a few people misunderstanding what 'mutating' means, biologically speaking. Mutation gets kind of a bad name in pop culture because it is often associated with significant deformations, or disease, or giant leaps in evolution. Mutation is a constant of biology, in that it is an ever present phenomenon. All biological entities, living or semi living in the case of viruses, mutate as they replicate. Mutation is simply a random change to DNA.
The question, evolutionarily, is what is the impact of that mutation? Most mutations are insignificant when it comes to organisms with large genomes with a lot of redundancy. A lot are detrimental. Very few are advantageous to the organism, and usually only happen to be based on the environment they're in.
This is a little different to viruses which have comparatively smaller and less complex genomes with little redundancy. Some viruses recombine their genomes to essentially induce mutations as one of their mechanisms of reproduction. Influenza does this, resulting in antigenic shift.
Antigens are molecules which induce an immune response, and which can cause the specific generation of antibodies (the things which bind to pathogens and neutralise them). Antigens are the things that our immune system 'remembers' when it develops immunity to a pathogen, and the thing that vaccines attempt to target.
Antigenic shift It is a random process that is one of the reasons influenza strains change so dramatically, so quickly, and why immunisations need to be seasonal. This is a sudden and dramatic change to the virus, and is a little different to how coronaviruses have worked, historically. Coronaviruses can have sudden and dramatic rearrangement of their genomes, but usually only as a result of coinfection in animals. Both coronaviruses and influenza are susceptible to something called antigenic drift, though, which is the ongoing minor changes to those antigens based on small random changes to the genome as a results of a 'proof reading error' during replication. Basically, the genome is replicated, and a small coding error is made, and not corrected. As an RNA virus, coronaviruses are particularly susceptible to these little changes, and usually have some continuous fluctuations in what antigens they express. This is why vaccines are so hard to make against these types of viruses. Covid19 has appeared relatively stable to date, but there's no reason for that to be the case throughout this spread.
These types of mutations are the really critical part here, because being able to reinfect hosts, or avoid an immune response is a massive selective advantage to the virus. That is, the mutation is random (changing RNA which changes antigens), but the outcome is beneficial, so it will be conserved and replicate in high numbers. This is different to random mutations which make viruses more or less lethal.
Mutations can also make a virus more deadly. They can also make it less lethal. Here's the thing though, being more lethal should be a selective disadvantage because being more lethal means the individual is quarantined, isolated, and potentially dies. Being dead is not a great way to pass on disease. Being alive and coughing a lot is much better for the virus. So the virus may mutate, but that doesn't mean, as someone here said, it's 'intelligent'. It's no more intelligent than a filter. The mutations that help it avoid the immune system increase in frequency, the mutations that cause it to be more difficult to pass on decrease in frequency.
This is, of course, a complex system and predictions are borderline impossible. For instance, hospitals may see particularly lethal strains in higher proportions, due to the selective event of the more lethal variants being hospitalised. This, along with viral load, may be one of the reasons so many healthcare workers are dying. And it may be that causing a really severe cough is enough of a reproductive advantage to the virus to outweigh the potential disadvantage of killing a host. Who knows. But it may also be the opposite. It may be that we see a less virulent strain spread across communities that effectively isolate symptomatic individuals.
This, by the way, is another fantastic reason for effective quarantine. If we quarantine and isolate anyone symptomatic, we introduce an artificial pressure (just as powerful) toward variants which are less lethal, whilst still creating some form of community immunity.
I'd like to reiterate that there are heaps of unknowns here. And it does seem that Covid19 is more genetically stable than we might otherwise expect meaning a lot of this is moot. At least in the short term. But I'd just like to clarify that the virus is not intelligently seeking to become more harmful. And that mutations aren't necessarily bad things. Like all things in biology, this is an evolutionary arms race, and like most evolutionary arms races, it will probably end up eventually in some form of balance.
