Picture a herd of elephants crossing the savannah. Each elephant carries 50 monkeys on its back. A sudden mutation makes it so that each elephant can now only

Picture a herd of elephants crossing the savannah. Each elephant carries 50 monkeys on its back. A sudden mutation makes it so that each elephant can now only carry 30 monkeys, but also doubles the number of elephants in the herd. This is terrible news for two-fifths of the monkeys currently riding. Some may never be able to climb back up an elephant and a few unfortunates may find themselves trampled underfoot. On the other hand, the monkey-carrying capacity of the herd is now 20% greater than it was before. Assuming we can sort out the logistics properly, there is enough room for all current monkeys and their offspring, plus maybe a few monkeys from outside who hadn't had a chance to ride yet. KEY: Fewer monkeys per elephant = displaced jobs due to automation. More elephants in the herd = economic growth due to automation. Monkeys = workers. Elephants = "the economy" or "the means of production". People who tell you that automation is going to lead to long-term net job losses from are betting on something that has a very drastic job-displacement effect (WAY fewer monkeys per elephant) while having a very small effect on economic growth (few if any additional elephants). That doesn't make sense to me. The implausibly perfect technologies they tend to talk about would create a massive amount of new value in the economy, i.e. a great deal more elephants for us monkeys to ride on, even if only four abreast. It's much easier to predict the jobs that will be displaced than the new jobs that will be created, so it's cognitively simpler to take the pessimist's side.