Throughout HPMOR, the author has included many fascinating details about how the real world works, and how to gain power. The Mirror of CEV seems like a lesson in what a true Friendly AI could look like and do.

I've got a weirder theory. (Roll for sanity...)

The entire story is plausible-deniability cover for explaining how to get the Law of Intention to work reliably.

(All quoted text is from HPMOR.)

This Mirror reflects itself perfectly and therefore its existence is absolutely stable. 

"This Mirror" is the Mind, or consciousness. The only thing a Mind can be sure of is that it is a Mind.

The Mirror's most characteristic power is to create alternate realms of existence, though these realms are only as large in size as what can be seen within the Mirror

A Mind's most characteristic power is to create alternate realms of existence, though these realms are only as large in size as what can be seen within the Mind.

Showing any person who steps before it an illusion of a world in which one of their desires has been fulfilled.

The final property upon which most tales agree, is that whatever the unknown means of commanding the Mirror - of that Key there are no plausible accounts - the Mirror's instructions cannot be shaped to react to individual people...the legends are unclear on what rules can be given, but I think it must have something to do with the Mirror's original intended use - it must have something to do with the deep desires and wishes arising from within the person.

More specifically, the Mirror shows a universe that obeys a consistent set of physical laws. From the set of all wish-fulfillment fantasies, it shows a universe that could actually plausibly exist.

It is known that people and other objects can be stored therein

Actors store other minds within their own Mind. Engineers store physical items within their Mind. The Mirror is a Mind.

the Mirror alone of all magics possesses a true moral orientation

The Mind alone of all the stuff that exists possesses a true moral orientation.

If that device had been completed, the story claimed, it would have become an absolutely stable existence that could withstand the channeling of unlimited magic in order to grant wishes. And also - this was said to be the vastly harder task - the device would somehow avert the inevitable catastrophes any sane person would expect to follow from that premise. 

An ideal Mind would grant wishes without creating catastrophes. Unfortunately, we're not quite ideal minds, even though we're pretty good.

Professor Quirrell made to walk away from the Mirrror, and seemed to halt just before reaching the point where the Mirror would no longer have reflected him, if it had been reflecting him.

My self-image can only go where it is reflected in my Mind. In other words, I can't imagine what it would be like to be a philosophical zombie.

Most powers of the Mirror are double-sided, according to legend. So you could banish what is on the other side of the Mirror instead. Send yourself, instead of me, into that frozen instant. If you wanted to, that is.

Let's interpret this scene: We've got a Mind/consciousness (the Mirror), we've got a self-image (Riddle) as well as the same spirit in a different self-image (Harry), and we've got a specific Extrapolated Volition instance in the mind (Dumbledore shown in the Mirror). This Extrapolated Volition instance is a consistent universe that could actually exist.

It sounds like the Process of the Timeless trap causes some Timeless Observer to choose one side of the Mirror as the real Universe, trapping the universe on the other side of the mirror in a frozen instant from the Timeless Observer's perspective.

The implication: the Mind has the power to choose which Universes it experiences from the set of all possible Universes extending from the current point.

All right, screw this nineteenth-century garbage. Reality wasn't atoms, it wasn't a set of tiny billiard balls bopping around. That was just another lie. The notion of atoms as little dots was just another convenient hallucination that people clung to because they didn't want to confront the inhumanly alien shape of the underlying reality. No wonder, then, that his attempts to Transfigure based on that hadn't worked. If he wanted power, he had to abandon his humanity, and force his thoughts to conform to the true math of quantum mechanics.

There were no particles, there were just clouds of amplitude in a multiparticle configuration space and what his brain fondly imagined to be an eraser was nothing except a gigantic factor in a wavefunction that happened to factorize, it didn't have a separate existence any more than there was a particular solid factor of 3 hidden inside the number 6, if his wand was capable of altering factors in an approximately factorizable wavefunction then it should damn well be able to alter the slightly smaller factor that Harry's brain visualized as a patch of material on the eraser -

Had to see the wand as enforcing a relation between separate past and future realities, instead of changing anything over time - but I did it, Hermione, I saw past the illusion of objects, and I bet there's not a single other wizard in the world who could have. 

This seems like another giant hint about magical powers.

"I had wondered if perhaps the Words of False Comprehension might be understandable to a student of Muggle science. Apparently not."

The author is disappointed that we don't get his hints. 

If the conscious mind was in reality a wish-granting machine, then how could I test this without going insane?

The Mirror of Perfect Reflection has power over what is reflected within it, and that power is said to be unchallengeable. But since the True Cloak of Invisibility produces a perfect absence of image, it should evade this principle rather than challenging it.

A method to test this seems to be to become aware of one's own ego-image (stand in front of the Mirror), vividly imagine a different ego-image without identifying with it (bring in a different personality containing the same Self under an Invisibility Cloak), suddenly switch ego-identification to the other personality (swap the Invisibility Cloak in less than a second), and then become distracted so the ego-switch becomes permanent (Dumbledore traps himself in the Mirror).

I can't think of a way to test this without sanity damage. Comments?

(Cross-posted on The Life You Can Save blog, the Intentional Insights blog, and the Effective Altruism Forum).

This will be mainly of interest to Effective Altruists

Effective Altruism does a terrific job of appealing to the head. There is no finer example than GiveWell’s meticulously researched and carefully detailed reports laying out the impact per dollar on giving to various charities. As a movement, we are at the cutting edge of what we can currently evaluate about the effectiveness of how we optimize QALYs, although of course much work remains to be done.

However, as seen in Tom Davidson’s recent piece, "EA's Image Problem," and my “Making Effective Altruism More Emotionally Appealing,” we currently do not do a very good job of appealing to the heart. We tend to forget Peter Singer’s famous quote that Effective Altruism “combines both the heart and the head.” When we try to pitch the EA movement to non-EAs, we focus on the head, not the heart.

Now, I can really empathize with this perspective. I am much more analytically oriented than the baseline, and I find this to be the case for EAs in general. Yet if we want to expand the EA movement, we can't fall into typical mind fallacy and assume that what worked to convince us will convince others who are less analytical and more emotionally oriented thinkers.

Otherwise, we leave huge sums of money on the table that otherwise could have gone to effective charities. For this reason, I and several others have started a nonprofit organization, Intentional Insights, dedicated to spreading rational thinking and effective altruism to a wide audience using effective marketing techniques. Exploring the field of EA organizations, I saw that The Life You Can Save already has some efforts to reach out to a broad audience, through its Charity Impact Calculator and its Giving Games, and actively promoted its efforts.

I was excited when Jon Behar, the COO & Director of Philanthropy Education at TLYCS, reached out to me and suggested collaborating on promoting EA to a broad audience using contemporary marketing methods that appeal to the heart. In a way, this is not surprising, as Peter Singer’s drowning child problem is essentially an effort to appeal to people’s hearts in a classroom setting. Using marketing methods that aim to reach a broad audience is a natural evolution of this insight.

Jon and I problem-solved how to spread Effective Altruism effectively, and came up with the idea of a catchphrase that we thought would appeal to people’s emotions well: “Be a Superdonor!” This catchphrase conveys in a short burst crucial information about Effective Altruism, namely that one can have the most powerful impact of one’s donations through giving to the charities that optimize QALYs for the most.

More importantly, it appeals to the heart well. Superdonor conveys the feeling of power – you can be super in your donations! Superdonor conveys an especially strong degree of generosity. Superdonor conveys a feeling of superiority, as in better than other donors. In other words, even if you donate less, if you donate more effectively, you can still be better than other donors by donating more effectively. This appeals to the “Keeping Up With the Joneses” effect, a powerful force in guiding our spending.

Just as importantly, “Be a Superdonor!” is easily shareable on social media, a vital component of modern marketing in the form of social proof. People get to show their pride and increase their social status by posting on their Facebook or Twitter how they are a Superdonor. This makes their friends curious about what it means to be a Superdonor, since that is an appealing and emotionally resonant phrase. Their friends check out their links, and get to find out about Effective Altruism. Of course, it is important that the link go to a very clear and emotionally exciting description of how one can be a Superdonor through donating.

Likewise, people should get credit for being a Superdonor through getting others to donate through sharing about it on social media, through talking about it to friends, through getting their friends to go to their local EA groups. Thus, we get the power of social affiliation, a crucial aspect of motivation, working on behalf of Effective Altruism. A particularly effective strategy for social affiliation here might be to combine “Be A Superdonor” with Giving Games, both the in-person version that TLYCS runs now and perhaps a web app version that helps create a virtual community setting conducive to social affiliation.

Now, some EAs might be concerned that the EA movement would lose its focus on the head through these efforts. I think that is a valid concern, and we need to be aware of the dangers here. We still need to put energy into the excellent efforts of GiveWell and other effective charity evaluators. We still need to be concerned with existential risk, even if it does not present us in the best light to external audiences.

Therefore, as part of the Superdonor efforts, we should develop compassionate strategies to educate emotionally-oriented newcomers about more esoteric aspects of Effective Altruism. For example, EA groups can have people who are specifically assigned as mentors for new members, who can help guide for their intellectual and emotional development alike. At the same time, we need to accept that some of those emotionally-oriented thinkers will not be interested in doing so.

This is quite fine, as long as we remember our goal of making the strongest impact on the world by optimizing QALYs through not leaving huge sums of money on the table. Consider the kind of benefit you can bring to the EA movement if you can channel the giving of emotionally-oriented thinkers toward effective charities. Moreover, think of the positive network effect of them getting their friends to donate to effective charities. Think of whether you can make a much bigger difference in doing the most good per energy of effort by focusing more of your own volunteering and giving on EA outreach in comparison to other EA-related activities. This is what inspired my own activities at Intentional Insights, and the recent shifts of the TLYCS toward effective outreach.

What are your thoughts about reaching out to more emotionally-oriented thinkers using these and other modern marketing strategies? If you support doing so, what do you think you can do personally to promote Effective Altruism effectively? Would love to hear your thoughts about it in comments below, and happy to talk to anyone who wants to engage with the Intentional Insights project: my email is gleb@intentionalinsights.org.