“Human-compatible AI” by Stuart Russell (Data Driven Paris)

At getAbstract, we summarize books* that help people understand the world and make it better. Whatever we select for our library has to excel in one or the other of these two core criteria:

Enlightening – You’ll learn things that will inform and improve your decisions.

Helpful – You’ll take-away practical advice that will help you get better at what you do.

We rate each piece of content on a scale of 1–10 with regard to these two core criteria. Our rating helps you sort the titles on your reading list from solid (5) to brilliant (10). Books we rate below 5 won’t be summarized. Here's what the ratings mean:

10 – Brilliant. A helpful and/or enlightening book that, in addition to meeting the highest standards in all pertinent aspects, stands out even among the best. Often an instant classic and must-read for everyone.
9 – Superb. A helpful and/or enlightening book that is extremely well rounded, has many strengths and no shortcomings worth mentioning.
8 – Very good. A helpful and/or enlightening book that has a substantial number of outstanding qualities without excelling across the board, e.g. presents the latest findings in a topical field and is written by a renowned expert but lacks a bit in style.
7 – Good. A helpful and/or enlightening book that combines two or more noteworthy strengths, e.g. contains uncommonly novel ideas and presents them in an engaging manner.
6 – Notable. A helpful and/or enlightening book that stands out by at least one aspect, e.g. is particularly well structured.
5 – Solid. A helpful and/or enlightening book, in spite of its obvious shortcomings. For instance, it may offer decent advice in some areas while being repetitive or unremarkable in others.

*getAbstract is summarizing much more than books. We look at every kind of content that may matter to our audience: books, but also articles, reports, videos and podcasts. What we say here about books applies to all formats we cover.