Alexei(Alexei Andreev)

Karma: 3,393

Alexei 2 Aug 2010 15:15 UTC
0 points
on: Two Truths and a Lie
A modification for this game: come up with a list of 3 (or more) propositions, which are randomly false/true. This way the predictive theory can fail or succeed in a more obvious way.

Alexei 2 Aug 2010 15:40 UTC
5 points
on: Welcome to Less Wrong!
Hello everyone!

I’ve been quietly lurking on this website for a while now, reading articles as fast as I can with much enthusiasm. I’ve read most of Eliezer’s genius posts and started to read through others’ posts now. I’ve came to this website when I learned about AI-in-a-box scenario. I am a 23 year old male. I have a B.S. in computer science. I like to design and program video games. My goal in life is to become financially independent and make games that help people improve themselves. I find the subject of rationality to be very interesting and helpful in general, though I have trouble seeing the application for the more scientific parts (bayes) of rationality in real life, since there is no tag attached to most events in life. I would like to pose a question to this community: do you think video games can help spread the message and the spirit of this website? What kind of video games will accomplish that? Would you be interested in working on a game or contributing to one in other ways (e.g. donations or play testing)? Or maybe instead of writing games I should just commit to S.I. and work on F.A.I.?

Alexei 6 Aug 2010 22:02 UTC
3 points
in reply to: gjm’s comment on: Don’t Revere The Bearer Of Good Info
Granted that god exists and cares about us and he can change the world, even in tiny aspects, it’s very likely god will use those small aspects as a base to create the perfect world (kind of like AI FOOM). It follows that any world where god has some kind of minimum control will converge to the perfect world. Given that we are not in the perfect world, we can assume god does not have the minimum level of control.

Alexei 13 Aug 2010 20:21 UTC
1 point
in reply to: DSimon’s comment on: The Threat of Cryonics
I don’t think this quiet answers jtold’s problem though. What about when all these cryogenically frozen people will be unfrozen? What about refreshing each generation / where each generation starts anew so to speak?

Alexei 1 Sep 2010 16:24 UTC
1 point
in reply to: timtyler’s comment on: Should I believe what the SIAI claims?
This is very dangerous thinking. There are many potential holes not covered in your essay. The problem with all these holes is that even the smallest one can potentially lead to the end of the universe. As Eliezer often mentions: the AI has to be mathematically rigorously proven to be friendly; there can’t be any room for guessing or hoping.

As an example, consider that to the AI moving to quiescent state will be akin to dying. (Consider somebody wanting to make you not want anything or force you to want something that you normally don’t.) I hope you don’t come reply with a “but we can do X”, because that would be another patch, and that’s exactly what we want to avoid. There is no getting around creating a solid proven mathematical definition of friendly.

Alexei 2 Sep 2010 0:57 UTC
0 points
in reply to: timtyler’s comment on: Should I believe what the SIAI claims?

is not to try and make the agent do something it doesn’t want to—but rather to make it want to do it in the first place.

I don’t think there is much different between the two. Either way you are modifying the agent’s behavior. If it doesn’t want it, it won’t have it.

The problem with off switches is that 1) it might not be guaranteed to work (AI changes its own code or prevents anyone from accessing/using the off switch), 2) it might not be guaranteed to work the way you want to. Unless you have formally proven that AI and all the possible modifications it can make to itself are safe, you can’t know for sure.

Alexei 2 Sep 2010 18:45 UTC
1 point
in reply to: timtyler’s comment on: Should I believe what the SIAI claims?
I don’t think we really have a disagreement here. If you are building a normal program to do whatever, then by all means, do your best and try to implement safety features. Any failure would most likely be local.

However! If we are talking about building AI, which will go through many iterations, will modify its own code, and will become super-intelligent, then for all our sakes I hope you will have mathematically proven that the AI is Friendly. Otherwise you are betting the fate of this world on a hunch. If you don’t agree with this point, I invite you to read Eliezer’s paper on AI risks.

Alexei 3 Sep 2010 20:53 UTC
0 points
in reply to: timtyler’s comment on: Should I believe what the SIAI claims?
In fact, there is nothing vague about definition of “friendly”. Eliezer wrote a lot on that topic and I invite you to look at his writing, e.g. the link I gave you earlier.

I agree that if someone is going to launch a self-improving AI, then we will need to preempt them with our own AI if our AI has a greater probability of being friendly. It all comes down to the expected value of our choices.

Alexei 4 Sep 2010 0:57 UTC
1 point
in reply to: timtyler’s comment on: Should I believe what the SIAI claims?
Ok, I might have been a bit overenthusiastic with how simple “friendly” aspect is, but here is a good attempt at describing what we want.

Alexei 7 Sep 2010 19:51 UTC
13 points
on: A “Failure to Evaluate Return-on-Time” Fallacy
I find myself doing something very similar when it comes to every day activities. Let’s say I am searching for my missing sock and I have three locations to visit: A, B, C, with probabilities of finding my sock there: 80%, 15%, 5%. For some reason I keep wanting to search C, then B, before A. In my mind this kind of goes like this “let’s make sure it’s in none of the weird places and then I can be very sure it’s in A and search it thoroughly”. Empirically, most often if I just started with A, I would have been successful...80% of the time.

But to answer OP more directly: I feel that in most cases it’s simple akrasia and the fear of succeeding.
What links here?
- Memetic Hazards in Videogames by jimrandomh (10 Sep 2010 2:22 UTC; 136 points)

Alexei 10 Sep 2010 21:44 UTC
3 points
on: Memetic Hazards in Videogames
Hmm, I’ve never really confused my life with in-game life before, but I wonder if I maybe do it on the subconscious level. An interesting note: when I tried playing Morrowind (for those who don’t know, that game has a huge open world with many huge areas that are there for no reason other than to add realism), I had a sort of paralysis, because I had to explore every room and open every door, but that’s simply impossible in that game.

Alexei 10 Sep 2010 21:56 UTC
6 points
in reply to: SilasBarta’s comment on: Memetic Hazards in Videogames
An interesting article on stealing bicycles here.

Alexei 14 Sep 2010 21:04 UTC
2 points
in reply to: Will_Newsome’s comment on: Self-Improvement or Shiny Distraction: Why Less Wrong is anti-Instrumental Rationality
This is interesting, because I am also going down the same path: creating a list of things I need to do (daily, every so often, and one in a while). What website do you use for dual n-back task. Let me (and others) know if it’s effect. Is there any good literature/research on it? Does it help?

Alexei 14 Sep 2010 21:24 UTC
7 points
in reply to: DSimon’s comment on: Self-Improvement or Shiny Distraction: Why Less Wrong is anti-Instrumental Rationality
I’ve done the same. When I stumbled upon lesswrong, it’s like I got sucked into a vortex. I just kept reading and reading and reading, until I essentially ran out of things to read. I didn’t care about voting or commenting, just reading. Then I realized there are other people here besides Eliezer and other posts too. I realized the community here is pretty interesting, so I decided to join in.

Alexei 26 Oct 2010 20:45 UTC
3 points
on: Activation Costs
Actually, the same day you posted your article, Jess_Riedel posted this article on Willpower..

Alexei 2 Nov 2010 21:01 UTC
7 points
in reply to: Eliezer Yudkowsky’s comment on: Harry Potter and the Methods of Rationality discussion thread, part 5
I bet the Patronus’ power isn’t very strong when it’s far away for its owner. It’s just strong enough to get the message across, but not enough to repel dementors.

Alexei 4 Nov 2010 21:56 UTC
−7 points
on: The Curve of Capability
You make some interesting points, but completely fumble on your conclusion about AI FOOM. Please read the sequences. They are interesting and they will make you rethink your position.

Alexei 1 Dec 2010 22:04 UTC
−8 points
on: Cheat codes

The general trick was to figure out a new, stable state to change my habits to, then use willpower for a week or two until I settle into that stable state.

How do you define a “stable state”? Also, using willpower is not as simple as you make it out to be.

I think this post lacks content and originality.

Alexei 11 Dec 2010 4:49 UTC
0 points
in reply to: ata’s comment on: Life-tracking application for android
May be...The interface looks different, but, regardless, it’s close enough to be used as an example. Thanks!

Alexei 11 Dec 2010 17:59 UTC
1 point
in reply to: ideclarecrockerrules’s comment on: Life-tracking application for android
Thank you for the thoughtful comment.
- Plugin feature would be really nice, but not the easiest feature to implement. I like your idea of using intents though, so I will look into it.
- Yes, adding mood tracker and, more generically, ordinal values tracker is on my TODO list.
- Widget sounds like a good idea, I’ll add it to my TODO list. (I assume you’ll want to just create a widget for one track/interval type.)
- Yes, I am planning to add CSV import/export option.
- I would love to add more analytical tools, but I can’t think of any that would be really useful. I would be interested to know what people are tracking and what exactly they are trying to see/measure.
I am thinking about adding a paid version with more features, but I doubt I’ll make enough money to really justify that. I was planning to have ads in the version I published, but I just couldn’t bring myself to do that to the users. I hate ads.