Personally I don't think generative AI will be a marketable skill the same way mobile development was a decade ago or web development two decades ago. Using an

Personally I don't think generative AI will be a marketable skill the same way mobile development was a decade ago or web development two decades ago. Using an existing generative AI model is relatively trivial, just an API call, and so it doesn't have the complexity needed to create barriers to entry that'll justify high wages. Training a generative AI model does have that complexity, but it also requires compute & data resources that are beyond the means of an indie developer or small startup, and so you pretty much have to work for a well-funded company that can afford oodles of servers and petabytes of data. That limits the pool of potential employers, which again keeps salaries down. I think it might be much more like the economics for say aerospace engineers or petroleum geologists, both of which are very difficult technical fields, but do not have the bargaining power to demand high wages the way software engineers do. Hell, even like embedded software engineers, which arguably is harder than web/mobile development but paid significantly less because there are fewer businesses competing for those developers. Robotics potentially has a lot of potential, particularly if standardized hardware platforms emerge that could put the hardware aspects within reach of a lot of startups. It definitely has the domain-specific complexity, and yet will likely be structured as lots of small firms for individual markets searching for a common skillset rather than a few big firms trying to hire a very specialized skllset.