@claychristensen makes a distinction between 3 kinds of innovation 1. Disruptive innovation -> Takes something that is expensive and highly complex and only available to a few, and makes it cheap and simple enough that a much larger set of people can use it. The PC vs the mainframe is a good example imo. Early PCs were less powerful than mainframes, but democratised the power of computing to millions of people who could never afford or operate a mainframe. Also, for a minute think about the new jobs it created (pun intended :P) 2. Efficiency innovation -> This is about producing the same thing for the same customers, but with fewer resources. For example, Toyota reduced the time required to assemble a car from ~60 days to ~2 days. This type of innovation improves productivity, but they can also reduce the number of jobs. The good thing about these is that they free up a lot of capital on balance sheets which would've been stuck up in inventory, working capital etc. 3. Sustaining innovation -> Sort of makes a good product better for exisiting customers. For example your Iphone Pro Max 1899 replacing IPhone 1898. It mostly replaces demand from old customers and satifies it with a marginally better product. It probably doesn't have a lot of effect apart from increasing revenues. Talking about the Iphone, let's go back to Jobs. All of these innovations have different impacts on the new jobs created in the economy. Disruptive innovation uses a lot of capital. You need to spend money on R&D, experimenting to create product then spend some more in selling and distributing it. Plus it takes a decade or more to get your money back and probably why it isn't as lucrative. Meanwhile as we discussed earlier Efficiency innovation does almost the opposite. It reduces jobs and gives you a quick return on your capital. In an ideal state, both of these innovations work in a loop. Efficiency innovation -> frees capital -> capital gets invested into disruptive innovation -> new markets + new jobs -> repeat. Clayton disses modern finance because they obviously favour efficiency innovations because investing in them often looks much better on paper than investing in disruption. (Clayton has a lot of jokes on Modern Finance and Ben Bernake lol) Back to AI, It is true that if only we use it to automate the work companies already do it is largely an efficiency innovation and we will lose jobs. But, thankfully for Scaling laws we will have "Intelligence on a meter" making it cheaper and more accessible. Also a lot of these AI capabilities were already being utilised by the largest companies and they have been profiting from them for years and climbing to the top of industries. @AndrewYNg in his TedTalk (from back when they were good) talks about an example about how A Target or Walmart can spend millions building systems to forecast demand, optimise inventory, personalise recommendations etc. but small groceries can't. We will soon be able to give the same capabilities to every small store. Similarly now that building software and accessing intelligence have become much cheaper, we will be deploying it everywhere. And when the efficiency of harnessing a resource increases and its costs fall, we often discover a huge amount of demand from people/markets who couldn't afford it before. (H/t, The Jevons Paradox of Knowledge Work) I really think, we're all going to be so much wealthier (both as consumers and as producers). The age of intelligence!