Language policy for this blog

349 words

See also: content policy for this blog.

I have a habit of using ableist slurs. I used to think those words weren’t ableist at all, or maybe that ableism in that sense wasn’t bad. I have learned to think otherwise, mainly from this series of blog posts. Those posts highlight one important reason for avoiding certain words: it’s lazy language.

 “stupid” isn’t constructive. You’re not criticizing… you’re just denigrating. The other person can’t learn anything from being told that their idea is “stupid”. Like I said before… it’s lazy. It elucidates absolutely nothing.

Instead, you could be constructive. You could say “that’s a bad idea, and here’s why” or “I think you made a big mistake there. You should have done this, instead” or “I don’t know if I like that choice. Here’s a better one.” Explain why you don’t like whatever it is, instead of just calling it “stupid”.

Don’t have the time to go into specifics? It’s still better to just say something like “nah, that’s a bad idea” or “you know what? No. I’m against that” and move on than to say “that’s stupid”. Even when you’re using it against ideas or actions or such, there’s still splash damage.

Avoiding certain words forces you to be more articulate. To be more constructive in your criticism. To apply a growth mindset towards the outgroup.

(I am NOT saying that intelligence is a completely non-existent property. Some people are mentally impaired in ways that chronically prevent them from functioning normally, and it is harmful to think that this is not the case. I am saying that for most people, most of their usage of those words is inappropriate and harms their communication.)

Not only is there a practical advantage; using inherent properties of people as slurs is not so nice. I don’t feel welcome if I hear people use “gay” as a slur, so I should probably not use ableist language either.

Want to try this as well? Do read a list of alternative words like this one.  My favourite default replacements are:

  • Stupid -> willfully ignorant/frustrating/obnoxious/unfathomable
  • Crazy -> fascinating/amazing/extreme
  • Lame -> bad/awful/inadequate

1. Evaluating the AI Safety cause area: a two-year plan

644 words

[Most of this agenda didn’t survive contact with the enemy. You can find the latest installment here.]

My pretending to have readers makes this blog a self-commitment scheme: if I write that I will do something, the imagined possibility of social shaming will make me more committed to actually doing the thing. That is why I’ll line out my plans for evaluating whether I believe the AI Safety cause area as perceived by EA’s is any good.

I used to think the arguments were pretty convincing, and I liked how it made my skills in math and CS super relevant for a morally important thing. But then I listened to Bostrom’s Superintelligence: Paths, Dangers, Strategies and found the book’s arguments thoroughly lacking. There is an asymmetry of passion in the AI risk circles where the believers are hyping the cause area while the non-believers don’t interact with the ideas at all.

The following step plan is based on the assumption that the LW-aligned view of most EA’s is misguided, but that there is important work to be done to make sure that what we are currently calling AI is used properly. Every step in the plan is meant to defend part of my beliefs, and I trust myself to notice when my argumentation is shaky. If my view changes, I should feel free to change the plan to look into my new views instead.

After completing a step, I will write a blog post describing the outcome and related thoughts. I will try to complete at least one step every two months so that the plan is done before you re-enter the job market after the end of your PhD. I don’t have to do the steps in the listed order.

  1. Make a plan listing the steps to think through the AIS cause.
  2. Uneducated guessing: directly oppose the astronomical waste argument. See if you can collect the necessary entropy to argue that, even if astronomical waste would be astronomically bad, supporting the AI Safety cause is still bad in expectation. Read at least what Bostrom, Beckstead and FRI have written on the topic, maybe more.
  3. Reread Superintelligence. Try to list the major points in the arguments and list objections. Does it actually argue for a Yudkowsian view, or is everyone misinterpreting it and does it secretly[actually] argue for a more mainstream view?
  4. Read various online sources to get more of an idea of what the mainstream view is among AIS EA’s. See if you can list research agendas on AIS different from MIRI’s.
  5. Argue that MIRI is not an effective charity regardless of the status of AIS as a field. I am not sure I truly believe this, but they seem so deeply incongruent with the standard academic practice that I should spend some time thinking about them. I kind of expect that this post will feel like punching down.
  6. I think Paul Christiano is a smart and serious person. He co-authored one of the best papers in your field of the past decade so he is not a crank. Read some of his writing on AIS to see if it holds up to scrutiny.
  7. Argue that average utilitarianism is superior to total utilitarianism, and astronomical waste cannot exist.
  8. Argue in more familiar terms why a paperclip maximizer wouldn’t act as some people fear they would.
  9. Argue that the capitalist’s alignment problem, insofar as it is meaningful and solvable, will be solved by the market.
  10. The social democrat’s alignment problem is a meaningful concept.
  11. Educated guessing: redo the entropy gathering from earlier, but now while having more knowledge.
  12. Cast mainstream “AI” related research in TCS in terms of the social democrat’s alignment problem. See what is out there at conferences like STOC/FOCS, COLT and NIPS. Check O’Neil’s Weapons of Math Destruction, Dwork et al’s line of work on algorithmic fairness, the work on learning non-discriminatory predictors, etc.

Search the literature!

182 wordsEvery once in a while, I neglect to remember my favourite quote:

Six months of research can save you an afternoon in the library.

This is from a plaque in the library at work. The original quote is by American chemist Frank Westheimer. The longer I have been doing research, the more I am seeing just how true this statement is.

How to use this insight in day-to-day life:

  1. Realise that an afternoon is both really short compared to six months, and really long compared to how long you typically browse the web looking for an answer.
  2. Be aware that, in an everyday question, you don’t know the right words to query the search engine. After your first couple tries in web search and scholarly search (the results of Google and of Google Scholar seem almost disjoint so do try both), try to acquire the right keywords through figuring out what field your question is in, or to generalize your question.
  3. Google Scholar has interesting results even on page 10. Results are not ordered in relevance as well as web search does.

Rules for writing this blog

629 wordsDear future me,

In this post, I will state some ground rules that you have to follow for blogging. Your sense of self-commitment should be enough to stick to those rules, but in case it is not, I will try to give reasons that should convince you.


The primary purpose of this blog is to get better at writing. You have to write things that are challenging to write. Here are some different suggestions:

  1. Commentary: write an opinion piece, review a book or movie, criticise a policy or organisation. The challenge in these is to achieve multiple ends simultaneously: get your opinion across, be fair and intellectually honest, and be nice.
  2. Things you don’t understand: try to see how smart you can sound while talking about things you have little clue about. Be willing to put your understanding of a topic in words, even if you feel like you don’t understand it very well.
  3. Different writing formats: look around for fun writing formats and try to emulate them. In particular, you should try Buzzfeed, Vice, Math with Bad Drawings and Slate Star Codex.
  4. Bridge a gap in knowledge and caring: communicating is hard when people don’t know of each other how much they know of a topic, and when you know a lot about something, it is easy to overestimate how anyone else knows. Find topics where you can practice this.

I know you like writing about your favourite science experiments and you want to share your favourite tofu recipe, but you should wait a bit for that, gain some more writing experience. Critically review your earlier writing on that [not on this blog], then you can write it anew. No sharing of content for the sake of sharing content, only write the stuff you feel you would learn a lot from. This blog is for you to write, not for visitors to read.

A secondary purpose of this blog is to get more insight into the things you read. How much does someone need to know to seem knowledgeable? Are articles typically too long or too short to really make their point? Does having to produce a lot of content negatively impact how interesting it is?

The Rules

    • Every post should satisfy at least one of the following:
      • have a topic different from all previous things you have read in your life
      • be about your personal feelings,
      • have content that people can disagree with. Be provocative. Does your desired topic not satisfy this rule? Push it further and further until it does. This criterium should overrule your desire for global consistency, though local consistency within a single post should still be aimed for.
  • At least one blog post should go up every weekend. Try to write it in that same weekend, though allowance is made for when you are abroad for work.
  • Every once in a while, take a critical look at some earlier posts and see where you can improve. There are probably online communities of people who do this for each other. Consider joining one of those.

Just in case you do get readers at some point, it might be good to preface some posts with a short note about the intended audience? Think about this for a bit.

Silent edits are allowed up to 7 days after posting, as is adding links to the new article in old articles. After these 7 days, edits are only allowed to be of the following form. Inconsiderate viewpoint. [edit: past me held problematic opinions, sorry for that.]


You should stick to these rules for at least one year. After that, I invite you to evaluate the results, change some of the rules where necessary, and keep to them for another year. Good luck.

Forever yours,

Past me