Join 📚 Josh Beckman's Highlights
A batch of the best highlights from what Josh's read, .
The reason why there's no music here, is this is a reverential statement. I want you to know I'm not joking. I want you to know I mean everything I'm saying. This is not hyperbole, this is: I - very much - I'm being sincere here.
- Right.
And by getting rid of the fun music, and the campiness, it's also why this shot is handheld.
- Yeah.
It's a level of intimacy. It's so you, the viewer, you're sitting right in front of me.
Casey Neistat's SECRET to Filmmaking
Digital Spaghetti
The way double descent is normally presented, increasing the number of model parameters can make performance worse before it gets better. But there is another even more shocking phenomenon called *data double descent*, where increasing the number of *training samples* can cause performance to get worse before it gets better. These two phenomena are essentially mirror images of each other. That’s because the explosion in test error depends on the ratio of parameters to training samples.
Double Descent in Human Learning
chris-said.io
You are approaching this like an established natural sciences field where old classics = good. This is not true for ML. ML is developing and evolving quickly.
Ask HN: What Are the Foundational Texts for Learning About AI/ML/NN?
ycombinator.com
...catch up on these, and many more highlights
Readwise