A relevant quote from City Observatory summarizing research and implications:
But very little private housing in the United States was originally built for low-income people. Instead, homes built for the middle or even upper classes gradually became cheaper as they aged, as people with high purchasing power moved into trendier, more modern homes in “better” neighborhoods. As higher income households move on, the now somewhat older homes or apartments they formerly occupied are sold or rented to people with more modest incomes.
This process is called “filtering.” While the evidence that filtering is a real phenomenon has been around for a long time—the core of nearly every American city contains neighborhoods with once-luxurious homes now occupied by people of modest incomes—the first study to provide a rigorous measure of how it happens was published only in 2013. In it, Stuart Rosenthal of Syracuse University uses nearly 40 years of data from the American Housing Survey to figure out the average pace of filtering across the country, and what makes housing filter more quickly in some places than others. [1]
[1] http://cityobservatory.org/what-filtering-can-and-cant-do/
As mentioned, NYC isn't perfect but it has a reasonably functioning democratically elected city government that at least tries to deal with major urban issues and occasionally succeeds. Which I've heard anecdotally at least makes it a little different from San Francisco.
So you decrease the square footage from 1400 to 700.
You have to have customers willing to buy your product, and there are only so many $3000/month apartments you can make before you run out of people willing to pay that much. Either large numbers of last year's luxury places are moving downmarket or you won't be able to sell the building for what you expected.
Add to that, most of these projects are going to be financed by debt of some sort, and lenders are not going to be keen on seeing that you could pursue a higher profit margin and choose not to.