1. Better Language Models and Their Implications Open AI trained a large-scale unsupervised text generator with 1.5 billion parameters on a dataset of 8 million web pages. The text generator adapts to the style and content of the conditioning text and apparently outperforms other language models trained on specific domains.
  2. Do Neural Nets Dream of Electric Hobbits? by Scott Alexander of Slate Star Codex. Alexander follows the Open AI news with a great response. Relatable quote: “This isn’t a fanfiction, this is a dream sequence. The only way it could be more obvious is if Aragorn was somehow also my high-school math teacher.”
  3. Physics has Demoted Mass by Jim Baggott for Nautilus. While a common understanding has lead us to believe that matter has energy, quantum physics has turned the tables: Mass is a behaviour of quantum fields and not an intrinsic property to them.
  4. Introduction to Named Entity Recognition by Suvro Banerjee. Great article on Named Entity Recognition (entity extraction), a vital tool in Natural Language Processing.
  5. Think of a Number: Why Humans and Machines Are Bad at Being Random by David Britland with I Am Going to Read Your Mind – Magic Trick by Zach King shows the power of common bias. “The universe might well be random, but magic tricks, like many choices we make, are not.”
  6. Deep Learning State of the Art (2019) – MIT lecture by Lex Fridman. Current status of NLP, deep RL, real-world applications and more. 
  7. Data Science is Different Now by Vicki Boykis. Great article on the changing demand and supply of data scientists and important things to consider when going on that road.
  8. A Tale of Two Systems: 19th Century Behavioral Insights From Poe, Austen, and Dickens by Michael Hallsworth and Elspeth Kirkman for Behavioral Scientist. There is a certain permanence in human nature, here delightfully exemplified by excerpts from famous novels. 
  9. The Cognitive Aristocracy by Nick Maggiulli for Of Dollars and Data. An educational system primed to identify gifted students has lead to an intellectual elite, discounting students with worse performance. Maggiulli proposes a sovereign wealth fund with equal ownership across citizens to counter inequality. 
  10. Knowing When Tech Does (and Doesn’t) Promote Positive Relationships by Jenna Clark for Behavioral Scientist. Recently I deleted my last social network account (Facebook) as its negative effects outgrew the positive. Clark gives a concise research-based summary of the circumstances that direct the results.