The bubble hasn’t popped yet. The reason --- managers love the idea of product cheapening, while unable to judge its quality.

In the context of American culture, the primary market of «AI», it is impolite to give negative feedback, especially not corporatively aligned. As short-term costs are reduced with (still) cheap «AI», bosses are happy to get themselves a raise. Focus is on numbers, not the long-term implications. Deeply prepaid-for media expectation, gives an easy argument against criticism: wait for the AGI in the next «few» years, it will only get better.

A consolidated («big») tech industry kills two birds with one stone: by lowering the prime cost of its services, they do support own narrative of significant salary returns of the «AI» (which they also happen to sell).

Managerial bonuses and short-term gains make the legend very believable, especially in the consolidated tech industry with not so many alternatives, and expensive migration. But layoff waves will eventually create niche companies, focused on quality LLM template simply won’t be able to match.

Industry consolidation, detachment of managers from the know-how of their workers, and poor understanding of quality, makes «AI» replacing employees very appealing. But only until real expert competitors start capitalizing on the slop of «AI»-first (cheap) companies, and frustration of their customers.