prior probability

In addition to our regular stream of scholarly readings (blog posts, book reviews, essays, law review articles, etc.), below are some of the books we are reading this month (updated 11/13):

1. Tyler Cowen, Stubborn Attachments (Stripe). This will be a re-read for us, as we read a previous (unpublished) edition of this manuscript two years ago or so.
2. Steven E. Landsburg, Can You Outsmart an Economist? (Mariner). The title of this book is a bit misleading, because the book ignores the entire subfield of behavioral economics. Still, it’s a fun read.
3. Cass R. Sunstein, The Cost-Benefit Revolution (MIT Press). We are going to read this book on the strength of Tyler Cowen’s recommendation. Our copy of this book just arrived (11/12), so expect regular updates soon.
4. Tom Wright & Bradley Hope, Billion Dollar Whale (Hachette). We just finished reading this tome, which tells an incredible tale of financial fraud on the most massive scale imaginable. Alas, the main culprits in this story (the Malaysian boy wonder Jho Low and the investment banker Tim Leissner of Goldman Sachs) are still at large!

For our part, we say, Good Riddance! Here are some reasons why:

1. Karissa Bell, “Why Facebook and Instagram [and Twitter] should kill the like button,” via Mashable.
2. Liraz Margalit, “Our obsession with ‘like’ (part 1),” via Psychology Today.
3. James Somers, “The like button ruined the Internet,” via The Atlantic.

But if Twitter were to decide to remove its heart-shaped like button, how would its algorithms work?

Happy Halloween!

Check out this report by Jonathan Marciano describing a fascinating experiment in which a group of corporate lawyers with decades of experience in business law competed against an artificial intelligence (AI) program called LawGeex. Their task was to spot issues in five Non-Disclosure Agreements (NDAs), a type of contract that is commonly used in many business deals. Here is a summary of the outcome of the experiment (emphasis added by us): “LawGeex AI achieve[d] an average 94% accuracy rate, higher than the lawyers who achieved an average rate of 85%. It took the lawyers an average of 92 minutes to complete the NDA issue spotting, compared to 26 seconds for the LawGeex AI. The longest time taken by a lawyer to complete the test was 156 minutes, and the shortest time was 51 minutes.” (See infographic below the fold. Full disclosure: It appears that the creators of “LawGeex AI” were the same ones who ran the experiment.)  Continue reading →