The link: Genesis-DB: a database for autonomous laboratory systems.

The pitch: An example of a database engineered specifically for large-scale automated laboratory systems, which require (or leverage) specialized approaches to data modeling, querying, etc.

Why read it: Large-scale autonomous laboratory systems in closed loop with AI active learning algorithms are going to be one of the next accelerators in chemistry, biology, and medicine (and if you're paying attention, they are already beginning to be). The existing literature and databases aren't very far from good enough (publication bias, bad quality results, suboptimal institutional incentives for experiment design, and a host of other problems). Paraphrasing the over-used quote, if I had five years to optimize a molecule, I'd spend the first four building a robot laboratory.

Why skip it: I don't necessarily like the specific engineering choices in this database (for historical and sociological, not technical reasons); I'd be somewhat surprised if this were the right evolutionary path after all. But that's a minor quibble.

TL;DR: AI-driven close(ish-)loop Large-scale autonomous laboratories are where AI rubber will really hit the road.