A place for machine learning conversations

Can Bluesky be what Twitter was?

Twitter was the foundation of my career. It’s where I published the first chapters of Interpretable Machine Learning, and it was the reason I stayed motivated to finish the book. Through the ML community on Twitter, I got feedback and encouragement to keep going chapter by chapter.

But right around the time I became a full-time writer (in 2022), Twitter started to deteriorate: Heavily suppressed reach when posting links, especially substack links. Lots of weird ads. Worse vibe. Pay-to-win. Force-fed Elon Musk content. More bait, less nuance.

Recently, Bluesky, a Twitter clone, has taken off in terms of user sign-ups and I was curious to try it out. And I have to say, Bluesky has early Twitter-like vibes. In a good way. And there is a thriving stats and ML community, with much more engagement and discussion (I have 1/10th of the followers there, but get more quality engagement). What’s a bit missing from Bluesky is the cutting-edge AI stuff, which you can see as a con (won’t learn about the latest LLMs) or pro (lack of “You’re falling behind if you are not using AI”).

Anyway, I recommend giving it a try, especially if you like the concept of Twitter in general, but have found X lacking lately.

To get started I recommend looking for “Starter Packs” where you can follow people with certain focuses (like machine learning). Find lists here.

Also, let’s connect! You can find me on Bluesky here: christophmolnar.bsky.social