More and more software does not necessarily require more and more developers to write it.
It has so far. We might have expected the massive gains in developer productivity over the last 70 years to have resulted in the demand for new developers slowing down at some point, but it hasn't. In a very typical way, more supply just gets eaten up by more demand - if at any point we can afford to have fewer developers, we'll just end up writing even more software with the excess capacity.
It's hard to imagine a point of diminishing returns on software, since it's so useful in basically every walk of life.
Recently there has been an explosive transformation from software as a user and productivity focused industry into an adtech one, which has increased demand temporarily.
> In a very typical way, more supply just gets eaten up by more demand
Sure, but increasing supply also means decreasing price by that same model. Decreasing price means less money for developers, meaning less people are incentivized to become developers, leading to equilibrium. Of course, that's a very simplistic model.
It doesn't replace programming, it _is_ programming. It's useful because it makes programming more accessible and in some cases the visual representation has benefits over text.
I see more and more professional fields and application interfaces adopting things like SQL, Python, Notebooks (Jupyter etc.), JavaScript etc. Excel is obviously huge. A ton of technicians that work with physical stuff are also writing code in embedded interfaces or on large screens even. Programming is not a thing only software developers do.
Think of it this way, even if it were possible to run an enterprise through lowcode tools, an enterprise would still hire an agency or consulting firm to build it, and so the decision of what tools to use still goes to the specialists.