In 2026 I'm reinstating my blog with a commitment to LLM-free writing. I'm calling this "solowriting"—fully producing a piece of writing without AI in ideation, structuring, writing or proof-reading. Here's why I think it matters...
If you know a little bit about algorithmic fairness, you know that group fairness metrics are a popular way to quantify the extent to which a given classifier is supposedly "fair". But there are qualitatively different ways in which you can improve how a classifier fares under such metrics...
I'm back from FAT* 2020 (which will from now on be known as FAccT, and held next year in Toronto). This is a good time to round up a few of the presentations I found interesting. Partial summaries and random thoughts below...
As we all know, the internet is an incredible body of knowledge. To put things in perspective: today, anyone with internet access has — at their fingertips — access to a body of knowledge that even the most well-positioned scholar of a mere century ago couldn't reach in a year...
If you have internet access and a general interest in AI, you've stumbled upon the recent story of Amazon trying to build a machine learning algorithm to automate its hiring process, only to ditch it upon discovering that it penalized resumes that include the term "women's" in it...