1. Integrating Behavior Change and Persuasive Design Theories into an Example Mobile Health Recommender System by Torkamaan et al. (2021). Successful health apps for preventive and curative health care need to adapt to the user, reaching them at their point in the treatment journey and providing them with personalized recommendation of Behavioral Interventions .  The paper provides an implementation guide for the design of behaviour change based on the theories in the field, as well as the integration into the recommendation process. 
  2. A New Link to an Old Model Could Crack the Mystery of Deep Learning by Anil Ananthaswamy for Quanta Magazine. Until recently it was unclear why ANNs, with their immense amount of parameters, were not prone to overfitting. New evidence suggests the deep neural net is close to a kernel machine, and will match the evolution of the function represented by the kernel machine. In hyper-space, the function follows a classic gradient descent — at least in an ideal setting.
  3. Ignorance About Menstruation Puts Female Athletes at Risk by Louisa Nicola at Neuro Athletics. The impact of the menstrual cycle on sports performance is often ignored — partly because a lot of sports research focuses on males. However, taking the stage of the menstrual cycle and the corresponding availability of sex hormones into account can reduce the risk of injuries and increase performance.
  4. Inspired: How to Create Tech Products Customers Love by Marty Cagan. Cagan posits that the best tech companies create products in a very different manner compared to how most companies create products. While a little dated, it has become a standard reference that presents a unified philosophy of what the job of a product manager is and how to do it well.
  5. Want to fix Britain’s broken trains? Look at Austria by Nicole Kobie at Wired Magazine. Austria has introduced a public transport ticket called Klimaticket (climate ticket), which allows people to travel across Austria for three euros per day (1,095 euros annually). Not only does this make public transport far more affordable, it will likely motivate Austrians to use public options for leisure activities and inland travel. (The upside of the common sunk cost fallacy)
  6. Why Baselines and Deploying Early are the Most Important Ingredients for Successful Machine Learning Projects by Tobias Sterbak for Expertlead.  The majority of machine learning projects (60-80%, by some accounts) don’t make it to production. To improve chances of success, it is important to include domain experts early on, so that they can shine light on the intricacies of their field. In addition, even ML-heavy projects benefit from the agile route: Start small, test, and integrate early.
  7. Working Backwards: Insights, Stories, and Secrets from Inside Amazon by Bill Carr and Colin Bryar. The two authors were part of shaping Amazon’s principles and practices and involved with many of the products we use today. Some of the “wisdom” seems bland, but that only shows how much of this culture has become standard for today’s companies.
  8. We Can Prevent Dementia by Shane O’Mara at Brain Pizza. Estimates predict that by 2050 there will be 152 million dementia cases worldwide. Roughly 40% of these could be prevented or delayed by targeting 12 modifiable risk factors — such as smoking (rather obvious) and hearing impairment (less obvious).
  9. What Robots Can Learn from Fish and Fancy Math by Robby Berman for Big Think. A tuna’s fin stiffness changes with swimming speed — a skill useful for underwater robotics. Researchers at the University of Virginia created a robot that mimics the varying rigidity of a fish’s tail and discovered the corresponding equation:: Stiffness increases with swimming speed squared.
  10. Jhanas and the Dark Room Problem by Scott Siskind at Astral Codex Ten. The Dark Room Problem posits that if the brain aimed to minimize prediction error, one would prefer a dark room to the messy world outside. Siskind adds some thoughts on Andrés Gómez Emilsson’s perspective that the bliss experienced by certain stimuli is a combination of regularity/predictability/symmetry and enough complexity/unpredictability to hold our attention.