We overestimate AIs impact in the short-term and underestimate it long-termSolitude is not loneliness + clear(er) communication (Issue #281)Published inThe Medium BlogSent as aNewsletter3 min readJust now--Are you worried that AI might someday take your current job?I am, a bit. I run the Content team at Medium, which means we do a lot of writing in various forms particularly writing marketing copy, one of the main use cases for LLMs like ChatGPT. But also? Like a lot of the claims around tech trends, it feels like hype. Who profits from replacing workers with AI? Big tech companies, for one, who are also building or investing in a lot of these tools, and thus behind a lot of the hype.(I also happen to believe that meaningful writing requires humanity but thats a conversation for a different newsletter.)So how real is the threat AI poses to human jobs? The U.S. Bureau of Labor Statistics just put out a report on how AI will impact long-term job growth, and their answer is: It depends! They predict occupations whose core tasks can be most easily replicated by GenAI in its current form (like medical transcriptionists and customer service representatives) are likely to be affected, but also that employment in professional, scientific and technical services will rise 10% over the next eight years. Although it is always possible that AI-induced productivity improvements will outweigh continued labor demand, the report laments, really wringing its hands, there is no clear evidence to support this conjecture.On Medium, writer Ignacio de Gregorio thinks theres no chance AI will be taking any jobs in 2025, but for a different reason: cost. He points out that reasoning models are still highly inefficient, energy-wise, and running them at the scale wed need to replace humans is still far too expensive to make sense.If thats cold comfort, maybe AI pioneer Joe Procopios thinking on which jobs AI could replace will cheer you. Its not how good you are at AI, Procopio writes, Its how good you are at everything AI shouldnt be doing. Which is a lot. For example, what makes a great salesperson or software engineer isnt about mastering rote tasks (which AI can do), its about making smart decisions in context. In other words, AI will bring with it new opportunities. Enrique Dans, writing about the same BLS report, agrees: technological innovation has never been a unilateral force of job destruction. Feeling better?I like Rita McGraths framing that the question about AI and jobs is likely to be another example of Amaras law, that we tend to overestimate the effect of a technology in the short term, and underestimate it in the long term.