The Future of From AI To ZI
For the past few months, the blog’s regular readers may have noticed its irregular writing. This is a short post to explain what I’ve been up to instead of writing blog posts for the past few months, and explain what to expect from the blog going forward (hint: not much).
As some of you may know, I was supported by a grant from Open Philanthropy from October 2022 to September 2023, and as my grant wound down I shifted my writing efforts from blog posts to papers. This resulted in two papers, on Sparse Autoencoders and Attention-Only Transformers (the later of which grew from these two blog posts). I’m proud of those papers, and I’ll be presenting at NeurIPS workshops this week! (Of course I love all my children equally, and I’m also quite fond of some of my other research projects that didn’t turn into papers, such as this research report and its followup.)
With my grant now behind me, I know that I want to keep working in AI safety. While I don’t have any exciting news to share yet, I’ve applied to work at several AI safety orgs and research groups and I promise I will boast of my triumph here as soon as I can. Side note: if you’re someone who wants to hire me or fund me to do research, feel free to reach out through my email on my about page! I also have some individual research projects that I’m working through at a leisurely pace.
So what can you expect from the blog in the future? Probably the next post will be scoring my predictions from this blog post, which I cunningly timed to all be about “the end of 2023”. After that, expect infrequent posts which may cover a wider range of topics than just AI. In the next step of my career I hope I’ll still be writing public communications about my AI safety research, but that may be through a more official channel than this blog. If that happens, the distribution of posts on this blog may shift even further to the ZI end of the spectrum.
Thanks to all my loyal readers, and I hope to be in your inbox again soon!