Hi, I’m Azeem Azhar. I convene Exponential View to help us understand how our societies and political economy will change under the force of rapidly accelerating technologies. Every Wednesday, I do this in Charts of the Week.

CHART OF THE WEEK

The number of AI papers published on Arxiv is growing exponentially with a doubling time of roughly 23 months. Mario Krenn and colleagues reckon that scientists could use AI tools to help them decide on potential directions for research.

There is a system-wide acceleration in the volume of AI research. I dig into how AI is entering a new phase in Monday’s member-only commentary.

Share this chart

Quantity or quality?
Including Chinese language papers, Chinese researchers are producing 4.5x as many AI papers as Americans, with a much stronger focus on image detection, surveillance and autonomous machine research.

Ok computer
A recent study on machine learning compute load shows that the demand for processing continues to increase. The researchers identify a post-deep learning era which they call the “large-scale” era. Compute demand for breakthrough systems is doubling every 9.9 months or so, a slightly lower clip than during the deep learning era.

A thought about the intelligence explosion. Ian Good thought of it as a single ultra-intelligent machine. But in some sense, our economy (comprising firms, research institutions and the economic incentives) is a machine. And it is getting into progressively higher gear around AI research. 

DEPT OF EYES IN THE SKY

The dramatic decline in satellite launch costs, optical and other sensors, satellite bandwidth and the machine vision and data processing to analyse earth observation data is creating new insights.