I've been obsessed with GPT-3 since the original paper came out and I was reading it and interrupting the people I'm locked down with saying, you gotta hear this sample, this idea. It is one of the most important things that has happened this year, and I can't think of any better place to start than Gwern Branwen's very thorough explainer.

I've also been noticing the weird cloud-cover "anti-racists" have for pretty explicitly racist stuff – some of it literal praise for Hitler – and haven't found a good way to improve the conversation. But if you want a more sane society, understanding is the first step, and Bryan Caplan has a decent model of the current state of things.

The 21st century will be defined, in no small part, by the fate of liberal democracy and freedom as China and India rise. India is an interesting test, to see if a wildly corrupt government can retain control of a highly fragmented society (the sheer volume of distinct languages is a good lens into the diversity there) without devolving to authoritarian rule. But the test China is putting the world to is more direct. How do you structure your values to be resistant to the ever-increasing influence of genocidal authoritarians? How does humanity's endowment of the future not get polluted by it? Matthew Yglesias has an answer, it seems like it'll be a very good one, and this is precisely the national conversation we need to have. Pre-ordering both helps week 1 sales numbers so it will be a best-seller (getting post-release media coverage), and encourages the publisher to invest in marketing (getting pre-release media coverage). So pre-order, if you can.

Nick Punt has some ideas on how to make Twitter stink less. I think these ideas are pretty good, and add a technical mechanism to enforce the etiquette of, when you want to talk about what someone said without talking to them you need a screenshot. But Twitter is still not a place for conversations, and this doesn't strike at the root causes at all. It's too short, too impersonal (voice over text seems like a good start), and too media-oriented (why should random people see your posts?). While all things should improve, I'm skeptical you'd want to actually admit a mistake on a platform as toxic as Twitter, so you might as well not use it at all, mistake-flagging-feature or no. Good reminder to use "ability to make a sincere apology, admit mistakes, change your mind for good reason, contemplate plausible ideas you haven't judged yet, etc." as a way of evaluating community discussion.

There's a lot of uncertainty in the American presidential election. There's tons of turmoil right now, and a lot could change between now and election day. But it's possible we are underestimating how weird it could get – in particular, a growing suspicion in public discourse that the President is a child rapist – that can only be understood watching the very strange and much-publicized full interview with Jonathan Swan. I'd prefer not to characterize what he said, when he truly speaks for himself, and accurate modelling of what is going on is sensitive to details.

If you're interested in learning physics, or at least building real footholds for when you actually do learn it, you're probably disappointed in pop science communication. This isn't about you, it's about them – but Sean Carroll has been doing an admirable job of bringing substance to a TED-level vapid industry. And don't feel weird about sitting down with a textbook, struggling with it, thinking through it, and doing practice problems. These books exist for a reason, and some of them are even quite good.

Some very smart people you should probably pay attention to generally (notably Anders Sandberg and Stuart Armstrong, but their coauthor Milan M. Ćirković seems thoughtful even if I'd never heard of him before) have said the reason we don't see any evidence of alien life in the visible universe is because, essentially, the relationship between entropy and information depends on temperature, so they're waiting for the universe to cool down so they can more effectively run their brain. The paper has a delightful Lovecraft reference in the title, and is quite readable.