Alex Altair

"That's not even on my radar's radar"

You can't whistle inside a didgeridoo??? It just stops happening

There are multiple selection pressures going on here. Angry people are maybe more likely to join social media, and social media makes people more angry. (Or maybe you implicitly meant "people on social media are angry", in which case, yeah.)

When I say something is aesthetic, what I mean is that I like it and I don't know why

Pro tip, you can even do it in the microwave

I find this to be emphatically not true for me! (Example tasks are debugging, dealing with bureaucracies, writing messages with careful social wordings.)

Do people have opinions on Kuhn's Structure of Scientific Revolutions? I'm doing research in a pre-paradigmatic field and I'm wondering if reading the book will help me research better.

Sorry, I just got here, it will take me a hot minute to get annoying

Wintry
Autumnal
Sumerian
Springy

As a consequence, I'm now doing AI alignment research full-time! It feels like the jury is still out on whether I am any good at this, but I enjoy it immensely. I have my own research mini-agenda in agent foundations, and I wake up every day with a million questions that I want to find answers to.

So, I now have people body-doubling with me 40 hours a week. This is a life-changing breakthrough, and it is also very expensive. I am very unhappy that I had to work for 10 years before I could afford to do the thing that makes me actually able to focus.

Normally, whether I can focus enough to do anything is a coinflip, and I had to arrange my entire life around this severe deficit. With body-doubling, I can do work. It's not a marginal improvement, like stimulants are for me; it's an on-switch for working. If I am body-doubling then I am working.

The second part was that half way through the year I tried out body-doubling. This is an attention-control technique where you have someone sit next to you while you do a task.

(I wrote about this in much more detail here; https://www.lesswrong.com/posts/rTJrqtDLxAPxiW3sk/my-first-year-in-ai-alignment )

At the beginning of 2022, I participated in the AGI Safety Fundamentals course, and it went really well for me. This had just enough structure to keep me engaged, and then by the end I had some level of special-interest energy in a topic, so I kept it rolling for a while.

The first was that I started to usefully engage with AI alignment. I has always been a big part of my life; I was immediately convinced by the arguments about existential risk from AI. But because of my Life-Long Attention Problem, I have never been able to be usefully productive towards it.

2022 was one of the best years of my life. This came in two parts.

Here is a life update thread for people who know me. <3