The whole argument starts with a definition of what "work" is, obviously.
For people that use a particle accelerator or make scrambled eggs work from home is nonsensical.
For people who produce computer code as their entire work output and rely on very sophisticated digital collaboration tools already work from home is mostly interchangeable with work from an office.
And then there's a really vast grey area in between. What about lawyers, or architects, or film and television executives, or whatever else you can imagine?
The answer is that it depends. But, even limiting ourselves to the subset of work where it's even plausible that it could all be distributed there are some important things to note:
A pretty relevant data point is that when there has been a genuine choice between fully remote and mostly in offices the mostly in offices approach was way more popular. To the point where Silicon Valley real estate spiraled out of control, or literally everyone with deep knowledge of cutting edge electronics manufacturing ended up in the Shenzhen region.
In any another context we'd call that a "market" and say pretty definitively that the invisible hand of that market decided that this is a more effective way to build companies.
It seems we're going to get another experiment, whereby many companies begin the experiment all over again. Perhaps something has changed, perhaps virtual technology has advanced to where this makes more sense. Perhaps the culture is at a tipping point.
Or, perhaps there's some advantage to people all being in the same place that will make companies that do it outcompete companies that don't.
It's quite plausible that this is what the market will decide, in the future. It's also a really important data point to note that it is not, in fact, what the market had decided up until now.