Is machine learning research worth it for now? [D]

The Paradox of ML Research

I am a scientist who just applied machine learning to my research (JEPA/Representation/Geometric branch), and it did wonders. It allowed me to see so many papers that I am still struggling to write up. From what I see, there are clearly a million possibilities not yet explored - for example, industrial data, patterns in nature, and more.

The Job Market Contradiction

Why is the job perspective so pessimistic? We clearly have unsolved problems, and for many, the potential of ML will be proven for sure. We also have money (according to the news). So why are jobs almost impossible?