I like this evidence review. Everyone in the space is reading everything @alexolegimas writes. As they should. But I think this is a good moment to say that I highly encourage folks to also pay attention to everything @FutureEconJacob puts out. He is quite prolific and quite sharp! Here is my off-the-cuff ledger: - Evidence of aggregate effects on employment count (e.g. change in unemployment rate): none - Evidence of aggregate effects on wages: none - Evidence of localized effects on employment count: some though disputed on the "AI-exposed early-career" subgroup - Evidence of localized effects on wages: none - Evidence of macro productivity gains: pretty thin on claims that TFP is up - Evidence of micro productivity gains: a lot, needs to be carefully interpreted (as Alex has written), but there is stuff showing up here. The gap between worker productivity and firm productivity seems important and explained by several compounding bottlenecks. - Evidence that work looks different: abundant The last category matters a lot to me. It is very plausible to me that economists broadly will do a bad job of producing research on frontier AI that coheres with what laypeople perceive. I imagine economists may fixate on familiar outcome metrics like employment counts, wages, and productivity. Simultaneously, workers may find their jobs, the labor market, and the overall conceptualization of work to be highly unstable In its worst form, I can imagine economists continuously repeating "nothing to see here" because nothing registers on specific historically influential economic statistic in a way that loses public trust. I worry this will be a stronger version of Solow's paradox.

We characterize Markov perfect equilibria (MPE): each player’s speed depends only on the current state (capabilities + safety) and is optimal given others’ strategies. [Qualifier ‘Markov’ is b/c strategies depend only on the current state; history-dependent equilibria later]

New post with the excellent @FutureEconJacob examining the evidence (so far) on the impact of AI on the labor market. We reviewed 20+ papers and datasets and attempted to synthesize the evidence. Our take: 1) There is no strong signal that AI has had a negative or positive impact on employment in the aggregate. 2) When looking at specific sectors and demographics, the evidence is fairly mixed. While some have argued for negative impact on AI-exposed, early career hiring, others have identified moderators such as remote work or macroeconomic trends. There is also some evidence of positive impacts as well. So has AI hit the labor market yet? When looking at the employment statistics--not really. But that does not mean that jobs are not changing significantly, and that there aren't substantial changes in what skills are becoming less relevant versus in greater demand. That's a topic for another post! Link: aleximas.substack.com/p/has-…

Has AI hit the labor market yet? @alexolegimas and my verdict after ~20 papers: not yet in aggregate. Unemployment and layoffs show almost nothing. But AI may already be cutting junior hiring in exposed white-collar jobs. Remote work may explain part of that decline. A 🧵

Has AI hit the labor market yet? @alexolegimas and my verdict after ~20 papers: not yet in aggregate. Unemployment and layoffs show almost nothing. But AI may already be cutting junior hiring in exposed white-collar jobs. Remote work may explain part of that decline. A 🧵

Has AI hit the labor market yet? @alexolegimas and my verdict after ~20 papers: not yet in aggregate. Unemployment and layoffs show almost nothing. But AI may already be cutting junior hiring in exposed white-collar jobs. Remote work may explain part of that decline. A 🧵