Inside TikTok's Killer Algorithm

- Sara Fischer tl;dr: New users are shown 8 popular videos, each representing a different trend or topic. Users and videos are clustered separately and, based on users' engagement, the user is recommended videos from the appropriate cluster. TikTok concede this creates "filter bubbles."

featured in #206

TikTok And The Sorting Hat

- Eugene Wei tl;dr: "A machine learning algorithm significantly responsive and accurate can pierce the veil of cultural ignorance." With half of TikTok's engineers focussed on its matching algorithm, Eugene discusses how the algorithm made the app.

featured in #203

Data Structures & Algorithms I Actually Used Working At Tech Companies

- Gergely Orosz tl;dr: "This article is a set of real-world examples where data structures like trees, graphs, and various algorithms were used in production."

featured in #193

Testifying At the Senate About A.I. - Selected Content On The Internet

- Stephan Wolfram tl;dr: Over the summer Stephan was asked by congress whether "algorithmic transparency" is a policy option with regard to regulating "persuasive internet platforms." He discusses the complexities involved and a couple of conceptual options.  

featured in #166

Biased Algorithms Are Easier to Fix Than Biased People

- Sendhil Mullainathan tl;dr: Bias exist in both algorithms and people. Bias caused by an algorithm is symptomatic of a poor data set. If we create a framework to improve data sets - to account for everyone - an algorithm can be trained properly to negate bias. Retraining people is harder. 

featured in #165

The Architect Of Modern Algorithms

tl;dr: Barbara won the Turing Award for developing the Liskov Substitution Principle. She discusses the harm of modern algorithms and her experiences of being an award winning female engineer. 

featured in #162