I am supremely interested in the plots Eliezer is running.

I posted a longer form of this as a review / solution. Here's a condensed version:

Partial Transfiguration works through a deep understanding of physics. It allows Harry to to create any physically valid state of the universe, as long as he can hold it in his mind.

What this means is that you don't need to Transfigure a gun in order to fire a bullet. You can just Transfigure a bullet in the state of having been fired.

This is what the ability to Transfigure any physically valid configuration really means. You don't need to make a bulky laser weapon. Just make a laser pulse: an arbitrary amount of high-energy photons, aimed in the right direction. Instead of a shaped explosive charge, make a shaped explosion. Instead of antimatter, make gamma rays. Instead of a black hole, dangerous to everybody near it, make a bunch of gravitons and aim them at your enemy.

So given all that, how should Harry kill his enemies?

Lasers are messy weapons. Even black robes are reflective in some wavelengths. Use too much energy and you'll get a fireball back in your face. Release the energy too quickly and it will create an explosion instead of steadily boiling away your target.

Kinetic energy is safer. Transfigure a set of diamond missiles-in-flight, one aimed at each Death Eaters and one also for Voldemort, who is conveniently floating behind them. Giving them a speed of, oh, 0.005c should do nicely. They should be as large as possible - in order to leave large holes - but, since the difficulty and length of Transfiguration scales with the size of the target form, they will be flat and thin: head-sized and a millimeter thick, lying on the ground in front of Harry until the moment when, Transfiguration completed, they instantaneously acquire the forward velocity (and some angular momentum) that will have them impacting the Death Eaters' masks a few microseconds later. The slight layer of air turned into plasma carried in front of them will serve as a nice bonus.

Transfiguring the ground in front of Harry, if possible, is the best solution. Lacking that, pieces of Harry's legs will serve. Since Transfiguration can change the size and mass of the subject, the resulting wounds need not be deep.

If Transfiguration scales with the diameter of the target form, rather than its volume, we will reluctantly use much smaller bullets. A thousand diamond squares, one centimeter across and a millimeter thick, will form a sheet 31 centimeters on a side: much smaller than a car battery. An average of 26 .40 caliber bullets per head should be sufficient to the task.

Harry can also dispel any of his transfiguration wandlessly and wordlessly. So any toxic substance he creates he can dispel as it reaches him.

Nice. Where did you find that? Either Uncle Google is failing me, or I am failing Uncle Google.

Is there any discussion on the uses of friendliness theory outside of AI?

My first thought was that It seems like it could be useful in governance in politics, corporations, and companies.

I heard about DAO's (decentralized autonomous organizations) which are weak AI's that can piggy back off of human general intelligence if designed correctly and thought that it would be useful for those things too especially because it has a lot of the same problems that good old fashioned AGI have.

Did you optimize your match score? 81% doesn't sound a lot. You can increase your visibility without reducing your honesty by applying the techniques of this guy:

http://www.wired.com/2014/01/how-to-hack-okcupid/all/

Note that you don't need to scratch OKC. Just apply adaptive boosts (weigh safe bets higher) and don't answer ambiguous questions or questions where you expect mismatches (except if they are important for you).

This brought me out of the ~70% range into the >95% range without any lying.

It sounds like mathematical platonism which appeals to some like (seemingly) Roger Penrose, but it seems connected to other networks of concepts at least in part and I do not think it should be taken as a given. Perhaps in the future we can model when such belief systems will arise given access to other information based on the history of the individual in questiin, but we are not quite there yet. For further reference one could examine the examples of computationally model led social behavior in a topic someone created here regarding the reverse engineering of belief systems

Started on some art projects recently... One is a portrait of my sister... another is some practice concept art for a game Idea centered around both regular and exotic explosions. The game may not ever come into existence,The portrait of my sister is more about practicing drawing people. Also, I don't know if this counts as a project, but I'm learning how to use DirectX so theres that. Another Item I'm not sure counts as a project because it came in a kit was a computer interface board I made that can be used in home automation stuff. I had to solder all the components onto the board and everything! hehe had fun doing it too. there is a switch on the board though that is acting like it's always on which suggests a short. I'll need to study the schematic and refresh myself on what the components actually do to debug it though. Will probably get on that next week when I have time.

All of these projects tie into practicing my world building skills.

Arbitrary, as in ones you pick yourself? Well, pick a problem, then Google it.

Do you mean random?

what is a reliable way of identifying arbitrary solved or unsolved problems??