If you compare the present to, say, the 1960s, there have been any number of subsequent technologies that have caused 'tectonic' improvements in software engineering productivity. High level programming languages, visual text editors, Intelisense, relational databases, spare RAM and CPU cycles, huge repositories of free libraries, extensive documentation published on the internet that anyone can access... There have also been technologies that have made it possible for non-engineers to perform complex computational tasks (Excel, and to a lesser extent SQL, being the obvious examples).
It could be that AI is going to have a bigger impact than all of that stuff. But my guess is that if that none of the above technologies caused the market for software engineers to crash, AI won't either. What we're experiencing at the moment is a regular market downturn that happens to coincide with peak LLM hype. It's possible to hire too many software engineers. A lot of companies did. In the long run, the amount of software that gets written will probably continue to increase with the ease of developing it. AI tooling will be one more productivity boost in a long line of such.