@Lou4Gardens At least in my work environment, I can use AI extensively when working independently. But when I have to explain my work to senior colleagues or clients, overreliance on AI becomes pretty obvious. So learning can be monitored and incentivized to some extent.
A fascinating result from David Autor et al.:
AI helps junior patent lawyers more than senior lawyers while they use it.
But after 3 months, take the AI away and only senior lawyers have gained expertise. Juniors show basically no average learning effect.
This paper by @davidautor & co-authors is one of the most important papers to date in the space of what AI does to learning.
It's highly credible. And very much consistent with my own experience with use of agentic AI in research and teaching.
https://t.co/R4PeyhrBER
@Lou4Gardens Yes. But firms can always track the work product and fire the workers that are not learning with AI. Once that incentive is created, I think everyone has more incentive to learn.
East Asian economic miracles are coming again! This time, instead of relying on exports, it is mainly driven by AI demand along the supply chain. Taiwan is the best example of that
@VarunGodbole@NateWitkin I see. You are saying that the effect of AI on junior workers can depend on how they get supervised, staffed and trained. This is usually embedded in the organization structure. Super important point!
@Chrisliu222 This is why I keep arguing education has to build thinking skills first, tools second. The seniors gained because they had an internal model to check against. The juniors had nothing to calibrate against.
@VarunGodbole@NateWitkin Great question! Law firms have internal market. Lawyers get staffed under a centralized system. They do not have a consistent supervisor for different deals. So i think org may not be a determining factor here
@Lou4Gardens I agree. But the paper also doesn’t show that treated juniors learn less. On average, they perform at least as well as the control group. What’s more interesting is the greater variance among treated juniors: some may become dependent on AI, while others may actually learn more.
AI can make juniors better at the job without making them better professionals.
That creates a serious apprenticeship problem: if AI does the work through which juniors used to learn, where do tomorrow’s experts come from?
@lianda_edu Rather than helping PhDs or juniors in this bad market, some of seniors even try to bully them. Their top publicaition is second tier but they demand full compliance on their advises on any interviewee’s paper who has already published in top 5.
@BrontosaurB@HannoLustig My point is that upgrading the industries is the best way to avoid competitions from below and still extract rents and profits
@BrontosaurB@HannoLustig I am sympathetic to your point. take US as example. It no longer exports cars or manufacture goods, instead it exports services and recently AI. However Europe still relies on cars and manu goods, which are exactly the target of Chinese govt industrial policies.
@BrontosaurB@HannoLustig The problem is what EU produces that Chinese people want. No innovation means dying export. But Chinese still want luxuries and vacations. But that is also substitutable as well.