I’m moderating an event on digital technology and productivity later this month, and Noah Smith just published a great column on the topic, based largely off of a new paper by Erik Brynjolfsson, Daniel Rock and Chad Syverson. Here’s a key bit:
Often, when a very versatile new technology comes along, it takes a while before businesses figure out how to use it effectively. Electricity, as economist Paul David has documented, is a classic example. Simply adding electric power to factories made them a bit better, but the real gains came when companies figured out that changing the configuration of factories would allow electricity to dramatically speed production.
Machine learning, Brynjolfsson et al. say, will be much the same. Because it’s such a general-purpose technology, companies will eventually find whole new ways of doing business that are built around it. On the production side, they’ll move beyond obvious things like driverless cars, and create new gadgets and services that we can only dream of. And machine learning will also lead to other new technologies, just as computer technology and the internet led to machine learning.
This is the idea behind Michael Hammer’s vision of “reengineering”:
It is time to stop paving the cow paths. Instead of embedding outdated processes in silicon and software, we should obliterate them and start over.
It’s also closely related to Rebecca Henderson’s idea of architectural innovation; she argues that incumbent firms are quite bad at changing their architecture in response to new technologies.
If you combine the power of reengineering with the idea that incumbents struggle to do it, you end up with something like Chris Dixon’s full-stack startup idea, which, sure enough, others are applying to machine learning companies.
All of which is to say, I find Brynjolfsson et al’s theory quite plausible. Machine learning will become more valuable as it is incorporated into how organizations are designed, rather than just inserted into current structures. (It’s also why I think that even if AI progress slows, business will still shift dramatically.)
That’s also why I think Brynjolfsson and Andrew McAfee’s latest book, Machine, Platform, Crowd is so valuable. Redesigning organizations isn’t just about machine learning; when you combine ML with crowdsourcing and other newer models, you end up with fundamentally different kinds of organizations, like Numerai:
Numerai is a hedge fund managed by an anonymous community of data scientists. It encrypts its data and allows anyone in the world to continuously apply machine intelligence to the set and anonymously submit price predictions back. Numerai turns these predictions into trades and compensates the best performing models with bitcoin.
One open question: how does the wild divergence of productivity between firms fit in, given that it’s driven largely by digital technology? Is it the case that the winners so far are the ones who’ve really organized around these technologies? Or are they just better at the lesser early adoption that barely moves the needle, and will be toppled by a new era of AI-full-stack startups?