Conor's Newsletter: Issue 43
Current state of machine learning, inefficient efficiency, and impactful experiments
Let's get to the stories and links that I thought were worth sharing this week. As always, enjoy the newsletter. If this was forwarded to you, you can go ahead and subscribe at conordewey.com.
What I read —
My favorite thing I read this week, this post from Vicki Boykis on the current state of practical machine learning is gold. Just read it. The Steam-Powered Days of Machine Learning.
Should you prioritize latency or throughput? By emphasizing latency we get feedback sooner. Most of us get this wrong. The coffee example used here is especially effective. Inefficient Efficiency.
Going back a few years, I just stumbled upon this summary of The Programmers' Stone, a theory and training course on how to think as an effective and adaptable programmer. The Programmers' Stone.
Elijah Meeks wrote about visualizing funnel data. This is a pain point that I've experienced, and have yet to find a satisfying answer to. We Live in a World of Funnels.
What I found interesting —
Qubit did a meta-analysis on online e-commerce experiments and the outcomes. Some of the most impactful changes might surprise you. What Works in Experiments.
Tons of good resources here if you're looking for more startup reading. This could keep you busy for way too long. Road to Scale.
This study on the effect of life events on satisfaction has been making its rounds. Pretty powerful visualization. Life Satisfaction.
Quote I'm pondering —
“Learn the rules like a pro, so you can break them like an artist.” ― Pablo Picasso