A lean chain of intent is necessary to hit a goal as the slack compounds each layer down the chain. This is terrible for exploratory action because the output space is myopically limited but necessary for goal setting. Fatter chains often produce more total value but that energy is dispersed and the revenue streams often cannot take advantage of the innovation.
An analogy I like to use is trebuchets and daggers. JavaScript ability is like a dagger, generically useful but fungible and always in some sort of demand. Trebuchets are much more complex and much better at doing certain types of damage but the only buyers are castle seiging kings who may have already invested everything in a catapult producing supply chain.
Most fields like archaeology or art history don't expect their specialization to be highly compensated but often fields that go through periods of hype don't realize their skills can lose relevance. There are a lot of people learning ML now for example but once commoditization of tools and techniques happens the actual number of jobs may become too low to have market liquidity. Unlike A/B testing consultants.
Currently everyone in software can have their cake and eat it too but it's a position of privilege and luck and often this is forgotten.