This has been cross-posed to eaforum, you can access it here.

 

ProjectLawful.com: Eliezer's latest story, gives readers entry-level policy experience.

For the surprisingly large proportion of us that didn't already know, Projectlawful.com is essentially HPMOR 2. Whenever someone in your life asks you half-jokingly asks "how can I become smart like you?", you no longer need to answer "Have you ever read Harry Potter?" because Projectlawful.com does not have Harry Potter in it.

Specifically, projectlawful is a resource I highly recommend for anyone considering changing careers in policy work or starting careers in AI policy. This is for three major reasons:

  1. It routinely references and describes the communication skills, strategic skills, and other necessary skills that indicate whether the reader is a good fit for policy work.
  2. Unlike modern education systems, it gives the reader a fair chance to "reset" their understanding of math to see what math really is, and gives the reader a second chance re-evaluate whether the reader's quantitative skills are high enough to choose technical alignment work instead.
  3. It makes rationality education entertaining. It takes self-improvement, intelligence augmentation, and problem-solving techniques and seamlessly integrates them into your slack/resting time. Effectively, it costs negative energy to read.

 

It lets you evaluate fit for AI Policy work

Lots of people considering AI policy are either too intimidated or not intimidated enough by the social skills required to survive the superficial rat race in DC. There's a lot of nuance there, especially since some roles (e.g. congressional staffing) do, in fact, require you to be a charismatic mastermind, whereas other roles e.g. think tank work give so many opportunities to substitute intelligence for social skills that basically anyone who takes rationality seriously can thrive (assuming that they try to do rationality right, rather than read a bunch of Dawkins and skepticism magazines).

Projectlawful spends the first dozens of hours of reading time in the nation of Cheliax, a DnD-esque satanist kingdom where most of the people worship the idea of being evil and exploiting others. Essentially, an amped-up version of DC, where everyone is a skilled liar and nobody cares about you or thinks much about you at all, no matter how much more insightful you are; but where everyone is always looking for an edge. Projectlawful describes, often in detail, exactly the kind of Goodharting that makes up the everyday bureaucratic life in DC.

Projectlawful also goes a long way to dispel the "total incompetence" model of bureaucracy and government, which many people in the rationality community mistakenly believe in due to cultural roots with libertarianism and academic economics. In reality, each bureaucracy has an extremely complex mix of competent people (defined by their ability to handle incompetence) and incompetent people (defined by their ability to Goodhart competence). And AI policy is certainly a policy realm where competent people and structures are encountered, at a much higher rate than other places in government. Projectlawful represents this by depicting the day-to-day machinations of competent people in Cheliax's government, as well as the miscommunications that are inevitable among all the liars in such wretched hives of scum and villainy as Cheliax and your local government.

 

It tells you whether you have an aptitude to enjoy math, and should work on Alignment instead of AI policy

The English-speaking education system tricked billions of people into hating math.

Realistically, if you dislike math, odds are good that you are one of those billions of people.

I remember back in 2014, when Harry James Potter Evans-Verres said "And no matter what, I'm not having anyone Obliviate everything I know about calculus. I'd sooner die". I definitely didn't sympathize with HPJEV there, since I was taking Calculus in high school at the time and I was willing to forget all of it for free.

In 2014, HPMOR failed to inform me that the Statistics 101 class I was about to take would be more valuable than everything else I learned combined from ages 14-18. Projectlawful does not make that same mistake (about math-in-general, not statistics).

The male lead in projectlawful is a dath ilani, and he comes from a world that figured out how successfully teach math to children, in such a way that they don't grow up to hate it (which is considered one of many prerequisites to teaching math to children at all). It's made completely clear, in not-so-subtle comparisons made throughout the story, that math education on Earth has probably robbed you of the ability to enjoy math and evaluate (for real) whether or not you're capable of becoming extremely skilled at it. Unlike HPMOR, which praises math but doesn't feel the need to prove that math is worth your time, projectlawful routinely describes how creatively understanding math makes you stronger. 

Aside from being one of those lucky people who had a very good math teacher (perhaps including your parents), projectlawful is the only thing I've yet encountered that can give anybody a fair chance to see if math is right for them; not if they are right for math.

 

It makes Rationality fun

This one is pretty straightforward. After a long day of work, you can kick back with projectlawful for a few hours, and then go to sleep. You can read projectlawful on the weekend. You can read projectlawful on vacation. It's rest and rejuvenation and recharging, and also offers much of the same brain-augmentation as reading The Sequences. It's also a fine addition to CFAR's recommendation of actively implementing rationality in your day-to-day life by developing the habit of rational problem-solving; you get to have brain-boosting in your rest time, just like what HPMOR offered (although only reading projectlawful in your spare time is a clear violation of the CFAR's habit-building mandate).

There's also plenty of people out there who simply need an HPMOR-like work to make rationality fun, as it's the only way to persuade or motivate them to get started on rationality.

 

TL;DR

Reading projectlawful.com can provide you or someone else with critical insight on what kind of skills you have an aptitude for, in order to maximize your personal contribution to solving Alignment in time. It is low-cost time- and effort-wise, since like HPMOR, it is meant to be enjoyed your down-time.

I like this image, since it clarifies that being good at social skills/charisma is evidence that you are a good fit for AI policy, whereas skill at math is evidence that you're a good fit for technical Alignment work. Being good at both or neither pushes you towards the uncertain middle, and greater skill at one OR the other is evidence that pushes you in a specific direction. Being uncertain about red but certain about blue is also evidence about which direction you should pick. but for a lot of people, it takes something like projectlawful to have a fair chance at evaluating just how red or blue you actually are.

-4

New Comment
9 comments, sorted by Click to highlight new comments since: Today at 5:52 PM

You say :

Whenever someone in your life asks you half-jokingly asks "how can I become smart like you?", you no longer need to answer "Have you ever read Harry Potter?" because Projectlawful.com does not have Harry Potter in it.

On the contrary, this is a work I strongly wouldn't recommend, and especially not to newcomers. It's highly sexualized, contains descriptions of awful torture and various other forms of extreme misconduct, has a bunch of weird fetish material that more or less immediately disqualifies it as an intro rec in my opinion (far more so than Harry Potter stuff), is very difficult to get into thanks to the formatting, and also just... generally isn't all that good? I like some of Eliezer's writing, but I think this is very much not him at his finest.

Further, I very seriously doubt the idea that reading about a fictional government ruled by hell is meaningfully providing any real policy experience at all.
 

I like some of Eliezer’s writing, but I think this is very much not him at his finest.

I like most of Eliezer’s writing, and I agree with this sentiment.

The big problem here is that this is a glowfic, and I simply cannot bring myself to read it in that format.

I understand that the glowfic format might be better for authors / creators, but it sucks for me, and (I posit) a lot of other people.

If they really want to make it HPMOR2, it's going to have to be cleaned up and presented in a different, more readable format. The standard book/chapter format was developed for a reason.

Well that sounds cool!

clicks link

Oh... it's in glowfic format.

After a long day of work, you can kick back with projectlawful for a few hours, and then go to sleep. You can read projectlawful on the weekend. You can read projectlawful on vacation. It's rest and rejuvenation and recharging ...

I did NOT find this to be the case – I found it way TOO engaging and that it therefore, e.g. actively disrupted my ability to go to sleep. I also found the story to be extremely upsetting, i.e. NOT restful or rejuvenating. As-of now, it's extremely bleak.

I very much DO like it and I am perfectly happy that it's a glowfic. (There are some mildly confusing parts, probably because of it being a glowifc, but nothing too bad.)

This is also the first story for which I viscerally felt the utility of providing a 'trigger warning'.

I don't know what in particular makes you think the story is useful as "policy experience". I'm skeptical that much 'real world' policy in any way resembles the story.

I thought it was an allegory for having to deal with evil superintelligences. I suppose, like The Matrix, you can interpret that as politics or as technology or even as theology... Anyway, if it does contain valuable lessons, someone should try summarizing them, for people who could benefit, but aren't going to plough through an isekai fiction.

I said someone should extract the valuable lessons. David Udell posted Dath Ilan's safety principles for a tool AI, that was interesting. 

I have slightly refined my understanding of the work. First, I didn't realize it was set in an existing fictional universe, an RPG called Pathfinder. Second, I think Eliezer wrote (tweeted?) that it's meant to be a work on "hard-core decision theory". So it's clearly HPMOR 2.0. The result is as if Bourbaki decided to write a play with Samuel Beckett - it's not something I would read for pleasure, but it's a work I might try to understand, because I'm interested in the author's oeuvre and intellectual trajectory, and because I might need to navigate it at some point. 

Dath Ilan is apparently a kind of high-IQ libertopia. Again, it's an interesting concept, but I have to say that a society based on economic calculation through and through, seems to be one in which extremely selfish personalities set the tone for everything. Not everything called selfish is bad - I am especially thinking of transhumanism; humanity's relative indifference to the prospect of radical life extension might be its most foolishly self-denying trait - but a culture in which exploitative acts are restrained only by rational self-interest is surely less than optimal. (But perhaps that's not the entirety of how Dath Ilan works.) 

I'm already loving the story!

One quick question: what kinds of resources might be out there that have more complete/comprehensive basics to becoming a mad investor? I mean there's good reasons you can't just ask "Where are the galaxy-brained secret quant formulas written down?", and I have a lot of the mental models already, but I don't think a normal economics textbook (or even much of the existing Sequences) set out anything like "here's a comprehensive list of the basic mental models to be using in the crossover between rationality and economics".

A few actions immediately suggest themselves:

A. Crack open one of the few "books/webpages that are just lists of generally-useful mental models" that exist.

B. Add some recommended particular textbook or website to my reading list, and then never get around to reading it.

C. Wait for adaptation into more Sequences.

D. Continue reading between all of this.

Right now I'm just doing (D) and considering doing (A).

EDIT: The closest thing to a "list of things that Keltham can pull out at anytime", might be this.

Wait this sounds unusually well-targeted to me. (I say this after you personally recommended it to me earlier; this post is tipping me much closer to "read it", when I get the time...)