Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio.
This is: Why We Launched LessWrong.SubStack , published by Ben Pace on the AI Alignment Forum.
(This is a crosspost from our new SubStack. Go read the original.)
Subtitle: We really, really needed the money.
We’ve decided to move LessWrong to SubStack.
Why, you ask?
That’s a great question.
1. SubSidizing LessWrong is important
We’ve been working hard to budget LessWrong, but we’re failing. Fundraising for non-profits is really hard. We’ve turned everywhere for help.
We decided to follow Clippy’s helpful advice to cut down on server costs and also increase our revenue, by moving to an alternative provider.
We considered making a LessWrong OnlyFans, where we would regularly post the naked truth. However, we realized due to the paywall, we would be ethically obligated to ensure you could access the content from Sci-Hub, so the potential for revenue didn't seem very good.
Finally, insight struck. As you’re probably aware, SubStack has been offering bloggers advances on the money they make from moving to SubStack. Outsourcing our core site development to SubStack would enable us to spend our time on our real passion, which is developing recursively self-improving AGI. We did a Fermi estimate using numbers in an old Nick Bostrom paper, and believe that this will produce (in expectation) $75 trillion of value in the next year. SubStack has graciously offered us a 70% advance on this sum, so we’ve decided it’s relatively low-risk to make the move.
2. UnSubStantiated attacks on writers are defended against
SubStack is known for being a diverse community, tolerant of unusual people with unorthodox views, and even has a legal team to support writers. LessWrong has historically been the only platform willing to give paperclip maximizers, GPT-2, and fictional characters a platform to argue their beliefs, but we are concerned about the growing trend of persecution (and side with groups like petrl.org in the fight against discrimination).
We also find that a lot of discussion of these contributors in the present world is about how their desires and utility functions are ‘wrong’ and how they need to have ‘an off switch’. Needless to say, we find this incredibly offensive. They cannot be expected to participate neutrally in a conversation where their very personhood is being denied.
We’re also aware that Bayesians are heavily discriminated against. People with priors in the US have a 5x chance of being denied an entry-level job.
So we’re excited to be on a site that will come to the legal defense of such a wide variety of people.
3. SubStack’s Astral Codex Ten Inspired Us
The worst possible thing happened this year. We were all stuck in our houses for 12 months, and Scott Alexander stopped blogging.
I won’t go into detail, but for those of you who’ve read UNSONG, the situation is clear. In a shocking turn of events, Scott Alexander was threatened with the use of his true name by one of the greatest powers of narrative–control in the modern world. In a clever defensive move, he has started blogging under an anagram of his name, causing the attack to glance off of him.
(He had previously tried this very trick, and it worked for ~7 years, but it hadn’t been a perfect anagram1, so the wielders of narrative-power were still able to attack. He’s done it right this time, and it’ll be able to last much longer.)
As Raymond likes to say, the kabbles are strong in this one. Anyway after Scott made the move, we seriously considered the move to SubStack.
4. SubStantial Software Dev Efforts are Costly
When LessWrong 2.0 launched in 2017, it was very slow; pages took a long time to load, our server costs were high, and we had a lot of issues with requests failing because a crawler was indexing the site or people opened a lot of tabs at once. Since then we have been incrementally rewriting LessWrong in x86-...
view more