Newsletter
Addictive Work and other things
January 24, 2025
Hello,
Here’s everything since my last little missive to you:
Excerpt: It’s very trendy to say stuff like ‘start your day by making your bed and something something life is better’. But this is usually some kind of comment about the value of small and simple acts in promoting a sense of order and discipline. I’m not so interested in that. I’m more interested in those small and simple acts that make you addicted to those acts. I like other things that people say are addictive, so this sounds much more my speed, when it comes to productivity.
Main idea: The neural reward circuit implies that small, rewarding tasks that share environmental context are going to be the most addictive, so break tasks into small steps that end in a clear good feeling and optimise for a shared environment.
Open source GPT-o1. DeepSeek R1 matches OpenAIs GPT-o1 for performance, but DeepSeek is an open source model. Didn’t take that long to devolve the software. Not super surprising, given that efficiency is now the trick, not scale. The money will flow into the best uses for the models now, like OpenAI’s Operator.
–
Thoughts on AI use-cases. Whole AI update is interesting, if fragmentary. See also the bit about use-case limitations, lower down. Also the bit about getting ChatGPT to talk dirty to you. Fun.
–
The number of exceptional people. Honestly, I didn’t read it. Just that there’s a research paper on this amused me.
–
On how 4chan is becoming a hub for deep reading of classic literature. Obviously the New Statesman is making this about how the left are doing education wrong, but the underlying story is very interesting. Where else is this happening? Is it analogous to BookTok? If anyone has a write up that’s more focused, please share.
–
Rationalist blog publishes article telling rationalists not to rationalise away the bad vibes some people give out. I love these guys. As I say elsewhere, they’re a bit too worried about bias in thinking. Their online hub is called ‘LessWrong’. They are very committed to making decisions absent emotions (or something, lots of definitional things are heavily litigated there). Now, I don’t know how representative this article is of Rationalists more broadly, because I don’t live there. But it’s what I’d expect from them. It really does feel like they’re just rediscovering all the stuff we already know, and one of these things is that you can rationalise anything. That doesn’t put it under any obligation to be true. Like ignoring bad vibes.
–
Benjamin Libet is famous for doing an experiment in the 80’s, predicting what button people would press using EEG (brain) measurements about half a second before they were conscious of the decision they’d made. Plenty of heated debate about method and interpretation, but the general idea is that if the decision is made before we’re aware of it, then maybe free will isn’t a thing. Now, we have AI transformer models predicting your brain activity five seconds into the future from just ~20 second samples. See also the cases against free will. But importantly, I’m still not really sure that a lack of free will has many practical consequences?
–
I hope you found something interesting.
You can find links to all my previous missives here.
Warm regards,