Hi friends!
There's a lot to unpack this week!
But first I wanted to tell you a bit about Ship 30 for 30 a writing challenge that I joined last week. For 30 days I must write an atomic essay (200 to 250 words) every day or lose 50 bucks.
I'm only one week in and it's been an awesome experience. My goal is to improve a bit every day so that by day 30 my writing shows clear improvements over day 1.
You can keep up with my progress in this Twitter thread.
And if you're looking to improve your writing you should sign up for the next challenge starting January!
Cool Finds 🤯
Zachary Labe, a researcher at Colorado State Department of Atmospheric Science, made a list of very useful charts to visualize the extent of ice in the Arctic. It is not a pretty sight. According to some estimates, there hasn't been so little ice coverage for more than 1,000 years. And in case you still doubt human-made climate change is affecting the Earth go watch Sir David Attenborough's latest documentary.
The award for the most interesting story of last week goes to my friend JoaquÃn who while living in Japan doing his doctorate introduced his friend Yasuji D'Gucci to the cuatro (a sort of four-string guitar popular in Venezuela) who fell so in love with the instrument and its rhythms that he not only mastered it but over the years became a world-renowned Luthier. His cuatros are the most beautiful and polished I've ever seen and listened to.
David Goggins’ interview with Joe Rogan is one of the most inspiring and raw episodes I've seen on the podcast. It is worth listening to it from start to finish. David has put himself through so much discomfort that if you don't hear it from himself you might think some of it impossible.
Nerd Corner 🤓
Last week a recent article on AI bias caught my attention because this issue has been discussed at length. I'm still amazed that such stories keep surprising us.
Already in 2016, Cathy O'Neil was warning us in her book Weapons of Mass Destruction about the dangers of letting statistical models (what the media routinely calls "AI") take over a lot of processes and decisions.
The problem with this, is that it takes a human to design, code, and train these statistical models. And humans—by our very own nature—are biased consciously and unconsciously in a variety of ways.
From my days training computer vision models, I recall an episode in which one of our state-of-the-art action recognition models kept failing specifically on women. The culprit? Our dataset had over 80% males because most engineers who acted as test subjects were male... So we had to re-balance the datasets to ensure accuracy across genders and races.
Rather than being surprised and outraged by the results these models throw at us, we should focus on fixing the issues.
How? By recognizing that as humans we have biases we might not even be aware of and that the models will likely require a lot more training data and testing that we realize.
Perhaps we'll see the day when unbiased computer programs will generate other unbiased computer programs, but until then we'll have to be extra careful to avoid encoding our flaws into our digital lives.
Books 📚
While I'm still working my way through Edison's biography (by mid-book the writing has become a bit boring, just a recollection of facts), my dear friend Orlando sent me a signed copy of Rich Roll's new book Voicing Change. It contains interviews with some of the most inspiring characters alive taken from his podcast. My favorite ones so far are the ones with Ryan Holiday and David Goggins (check out his interview I linked above).
Before you go 😎
If you’re enjoying these emails it would mean the world to me if you share it with a friend or two. Just send them this link to sign up. I’m trying to make this one of your must-read emails each week!
And, if you come across anything interesting this week, send it my way. I love discovering new things through Nerd Corner readers!
Have an awesome week!
Alberto