Welcome to Things we read this week, a weekly post featuring articles from around the internet recommended by BMJ Labs.

Publishing

• Kent Anderson writes about  Why digital publishing is far more expensive than print. With a nice comment from Toby Green to highlight the point:
  “To give some numbers, at OECD we publish around 300 books annually, therefore we used to handle 300 ‘objects’ through our editorial, production and distribution process. Today, in addition, because of granularisation and multiple e-editions, we’re handling over 40,000 individual digital objects. Digital = complexity = cost!”
• The Times Higher Education wonders if more journals will follow suit as PNAS calls time on print after 103 years and what the future of journal publishing will look like in the digital age.
• eLife’s trial of a new approach to peer review proves popular with authors. Results showed similar acceptance rates for male and female last authors, but with higher acceptance rates for late-career researchers compared to their early- and mid-career colleagues.
• Forbes is building more AI tools for its reporters including a new CMS, called Bertie, which recommends article topics for contributors based on their previous output and headlines based on the sentiment of their pieces. They’re also testing a tool that writes rough versions of articles that contributors can polish up.
• Did you know that DOAJ’s metadata refreshes every hour? Nope neither did we, more on the DOAJ blog.
• Nature News report on what scientists searched for in 2018: AI is up, stress is down:
  

And finally…

We liked this infographic of The 20 Internet Giants That Rule the Web showing how much has changed over the last 20 years. (This ranking uses ComScore data which is focused on the U.S. and looks at unique visitors/viewers). Our guess is that a similar chart for academic publishing would show surprisingly little change.