Happy Wednesday!

This Chart of the Week looks at a chart from Our World in Data.

Ever wondered how computation resources used in training AI models has changed over time? Well, wonder no more!

There has been a massive increase in the amount of computation resources needed to train AI models. Language models top the list as the most hungry models needing more than 100 million petaFlops of compute resources for training.

1. Do you think this is sustainable going forward?
2. Any ideas on how we can build powerful AI models with fewer computing resources?

Message amyflorida626 if you would like to share a chart of the week via Zindi.