Mastery, Recommendation Algorithms, and AI Fine Tuning
Everything we published this week!
Hello and happy Sunday!
Want to learn to make AI write in your voice or the voice of your favorite writer? Come check out our live workshop with Dan on Wednesday, March 15th at 12 PM EST.
Now, on to the articles!
Evan Armstrong / Napkin Math
Recommendation algorithms have a huge impact on our lives. But how do they work—and are they good?
In this week's essay, Evan examines the architecture of recommendation algorithms and unpacks the tradeoffs tech companies are forced to face in building them. It's clear that no recommendation algorithm is perfect—but are they a necessary evil? Read on to find out.
People are building AI-driven copilots for developers, designers, and more—even Dan has been building his own.
In this week's Chain of Thought essay, he shares a simple checklist for builders looking to create their own AI-driven copilot. He unpacks the pitfalls of building copilots with current technology, and helps you understand where they'll be useful—and where they'll need time before they'll work.
Nathan Baschez / Divinations
We all want to achieve ambitious goals—but how do we stay calm and spacious along the way?
In this essay, Nathan explains how Robert Greene's book Mastery has helped learn to care less about your results, and more about the process of creating great work. And he shares one daily ritual that's allowing him the space he needs to stay on track.
Why do so many of us climb career ladders that we don't truly want to be on?
Simone Stolzoff borrows from the philosophy of games to explain why this happens so often—and how we can learn to play the career game differently
This week, Twitter broke for several hours: Links weren’t working, images disappeared, and TweetDeck went down. Twitter has been a huge part of our world for many years, almost in spite of itself. But the product is now fraying—will a drop in users follow?
So far, the answer seems to be no. But that might change soon. In this essay, Gareth Edwards argues that you can degrade the experience of successful consumer products for some period of time without seeing any corresponding drop in growth or retention rate. However, if you push things too far you’ll pass a critical threshold—a trust thermocline—where usage collapses.
Is Twitter close to breaching a trust thermocline? Decide for yourself.
That's all for this week!