As an extension of this: LLMs lower the barrier to entry for more complex challenges, which could allow more average coders to effectively produce the same results as senior engineers as the models evolve.

In this scenario employers will have less incentive to hire as many expensive senior engineers when average ones with AI can meet the KPIs for much less cost.

This obviously makes assumptions about how good generative AI will get (it's absolutely not at this level yet) but they're reasonable assumptions given the evolution we've already seen in a short period of time.

It's also logical to assume that employers will seek to lower staff costs wherever they can, especially large publicly traded companies answering to Wall Street analysts.