AndrewKemendo

Karma: 124

AndrewKemendo Sep 2, 2009, 1:35 AM
3 points
in reply to: CannibalSmith’s comment on: Open Thread: September 2009
I would disagree. The overwhelming majority of the average human’s life is spent peacefully. It is actually fairly remarkable how rarely we have significant conflict, especially considering the relatively overcrowded places that humans live. Not to mention that it is only a small proportion of the human population that engages other humans destructively (not by proxy).

AndrewKemendo Sep 1, 2009, 12:51 PM
1 point
on: Open Thread: September 2009
In reading the Singularity Institute’s research goals, and the ruminations of Yudkowski, Wei Dai, Nesov et al. in postings here, the approach to developing friendly AI which stands out the most, and from my perspective seems to just always have been the case, seems to be exclusively logic based in the vein of John McCarthy.

I am wondering how the decision was made to focus research for SIAI on the pure logic side, rather than, for example building a synthetic conscious which uses the brain as a model?

To be sure, nearly all AI approaches overlap at some point. However it seems a bit like reinventing the wheel from an almost purely theoretical perspective, when you have a very good model to work from.

AndrewKemendo Aug 26, 2009, 9:08 AM
2 points
on: How does an infovore manage information overload?
My job is to read AND understand information such that I can repeat it back clearly and precisely, often succinctly from 7:30-4:30. 99.9% of this information comes off of computer derived sources. When I am home, at lunch and in the evenings (when I am not exercising or eating) I am either on the computer reading more or reading a hard copy book. On an average day I probably, between work and home, read through 250-300 articles or studies. In general I have about 80-90% retention with comprehension nearly complete—some of the more abstract things obviously take longer.

Here are my tips:
- If you use the same sources on a daily basis (like LW/OB) then create a hierarchy of visits. I find Firefox really useful for this with the bookmark toolbar and RSS feed—that way I can go left to right sequentially and see what I have read already. That way if I only have a short period of time, I get the most out of it.
- Open only the articles which you are very interested in, and open them all at once so that you can see them in your tabs. A word of caution though, don’t necessarily read them in the order that you open them. There is the tendency to read in order, however I have found that if there is an article I am really excited about, then that is the only one I will think about when I am reading through another one. This method allows you to take your work in chunks rather than burning through one particular source all at once.
- Continually update your hierarchy of sources, refining which give you the best information
- If you use aggregators such as reddit—read the comments. Often they distill or give better context than many articles. Contributing to the comments is also very helpful as you can ask questions and get the most out of the information you just read.
- Give yourself a break. After a half dozen or dozen articles, depending on their weight, go look at something completely non-demanding or get away from the computer all together. I have found that over the years, with increasingly complex subjects my “breaks” become less and less complex eg: after reading two PLOS studies I take a 10 minute 4chan break.
- Get rid of distractions. That includes Instant Messengers, open programs and other things which draw you away from the text. I have found that if I get pulled out in the middle of a story or article, my comprehension drops dramatically.
- If there are parts of whatever you are reading that you don’t understand—look them up immediately. This is a huge part of the comprehension aspect. Yes, it will slow you down considerably. However once having a good understanding, you will have better context for the rest of the article and can better frame whatever fact you just learned.
- Lastly. Don’t worry about your eyes going bad. They wont.

AndrewKemendo Aug 26, 2009, 8:41 AM
0 points
in reply to: billswift’s comment on: How does an infovore manage information overload?
Likely to have it wherever you go or in case of a computer meltdown. I know the biggest loss for me when my computer died not long ago was my bookmarks and tool bars.

AndrewKemendo Aug 23, 2009, 12:18 PM
5 points
on: The Journal of (Failed) Replication Studies

If regular journals don’t have good incentives to publish “mere” (failed) replication studies, why not create a journal that would be dedicated entirely to them?

There are a few Journals of negative results already out there:

http://www.jnrbm.com/

http://www.jnr-eeb.org/index.php/jnr

AndrewKemendo Aug 9, 2009, 1:51 AM
3 points
on: Would Your Real Preferences Please Stand Up?

So why prefer one theory over the other?

I doubt it is necessary to need to choose one or the other. I would imagine that both theories could be well represented in the population.

Conscious minds are potentially rational, informed by morality, and qualia-laden. Unconscious minds aren’t, so who cares what they think?

Are you suggesting that unconscious motivation cannot be changed by the “rational” conscious? I would argue that any significant behavioral modification is doing exactly this. One example that I think demonstrates this well is martial arts training. In the beginning the trainee unconsciously flinches and turns away from an oncoming threat while the trained expert tracks and deflects the threat unconsciously.

AndrewKemendo Aug 9, 2009, 1:00 AM
7 points
in reply to: Bo102010’s comment on: Dreams with Damaged Priors

I recall recently “waking up” to a burglar going through my closet, and I being unable to move anything but my eyelids. I started to get a little excited, but then I considered “How likely is it that a burglar silently defeated my deadbolt AND I spontaneously became paralyzed?”

Similarly enough, a few years back I was “woken up” in the same manner to the sound of breaking glass outside of my window. In the half awake state, I worked through the probability that someone was actually breaking into my car outside of my window and decided that if indeed this was the case and I wasn’t just dreaming I would hear further evidence (door closing, car starting etc...) and then take action. I did not and so I went back to sleep.

My prior probability was low that my car would get broken into. The evidence should have updated it, however I erred on the side of the prior and came to the conclusion that I was just dreaming.

Wrong. Sure enough the next morning my car had a broken window and my CD player and CD’s were gone (Car thieves like Tchaikovsky and Art of Noise apparently). I am a much lighter sleeper now.

AndrewKemendo Aug 7, 2009, 10:57 AM
19 points
on: Why Real Men Wear Pink

the people too poor to immediately replace all their clothes are wearing any other color.

This one hits close to home.

During my middle school years I was going to a private school with wealthy children. It was all my parents could to to afford regular clothes, however I wanted to keep up. So I saved my money and waited months to go buy the hippest shirt around. So I stroll into a dance that we were having with my (no joke) purple superman t-shirt and quickly get rebuked and laughed at. Needless to say, I was behind the curve on that one—and it never saw the light of day again.

False signal failed.

AndrewKemendo Aug 3, 2009, 6:57 AM
0 points
on: Pain
I don’t think anything about pain as a concept or feedback process is bad. The things which may cause or accompany pain however are bad or challenging and can cause harm to sustainable life. That is what people actually don’t like. I am unfamiliar with cases where pain does not accompany some destructive situation.

On the flip side, I like the pain in my muscles after exercise because I understand it’s origins and it is signaling to me that I successfully broke muscle down which I have a reasonable expectation of growing back stronger.

Pain in and of itself is the most effective signal for motivation that can occur in the animal kingdom. What we refer to as pain is simply a feedback mechanism for indications of change in homeostasis.

I wrote this blog post a while back on brain pain and how I think it is similar to muscle pain, which I think is apropos.

AndrewKemendo Aug 2, 2009, 10:26 AM
2 points
on: Open Thread: August 2009
I would be interested to see a top level post in which the community agrees on specific heuristics which are more detrimental than others and as a result more important to eliminate.

For example: The community would agree to a high agreement level that confirmation bias would significantly help their lives to eliminate, while the community may not agree to the same level on the necessity to eliminate the information bias.

This would help narrow down the more significant biases such that we could focus on tests and games which would help us eliminate these biases, similar to what Will_Euler is suggesting.

AndrewKemendo Jul 8, 2009, 10:16 AM
0 points
on: Avoiding Failure: Fallacy Finding

Take a video from a genre that generally has a high fallacy per minute ratio, and edit the video to play a klaxon sound whenever someone commits a logical fallacy or gets a fact wrong, followed by the name of the fallacy they committed flashing on screen.

How fortuitous! I have been thinking that this would be a great substitute for the current lineup of political television programs on the air now. A series of clips from the day or week that expose the fallacies and logical errors that politicians and blowhards make on a day to day basis. I know Politifact and Factcheck do a great job at calling these people to the carpet.

I wonder however, if the median viewer has a deep enough understanding of fallacies for a program like this to be effective or if it would take a host to explain each one in detail for the viewer not to get lost.

AndrewKemendo Jul 5, 2009, 12:08 PM
0 points
on: Not Technically Lying
I know that this is simply an exercise but I think it points out my epistemic belief that there are no such things as gray areas in deceit.

The doctor could have easily have said “It’s the weakest painkiller I have,” or “It’s the strongest sedative I have,” or any other number of technically true but misleading statements

In reality the Saline IV is neither a painkiller nor a sedative whatsoever therefore:

The strongest painkiller I have

is actually as much of a lie as the morphine statement.

Now of course you go on to say this:

Because language does not specify single, clear meanings...

However in this case there is a clear meaning for “painkiller” or “sedative.”

I think there are good examples of shaping perceptions (which is what NTL’s really seem to be) on these types of issues. Here is an alternative for your example. The doctor could have said this:

“This is something which might help your pain”

That is both technically true and verifiable as not NTL. Does it have the same effect as making a definitive statement? Perhaps not, but it also does not lead to a false conclusion about morphine or the ethics of the doctor which the patient may make when he is slicing and dicing.

If lying is never justified because of its effect on the listener, then neither is NTL.

To which I would agree, but I pose another question: Is shaping another person’s perception in any way, based on reasonable knowledge of the outcomes and given authority (Such as the case with doctors) a positive heuristic?

AndrewKemendo Jun 28, 2009, 11:43 AM
0 points
on: The Great Brain is Located Externally
My father told me long ago that intelligence isn’t knowing the answer—it is knowing where to find the answer. After all of my education and learning I find this to be a very true statement.

It’s not clear to me whether this means that we should be alarmed and seek to hone our factual memories… or whether we should devote our attention to honing our Google-fu, as our minds gradually become server-side operations.

I do analysis for a living and am constantly asked on the spot detailed questions about very specific things, so I have to hone those memories. However when I was learning how to do this better (through associations and number games) I realized that I was simply being a human google search.

In my profession, one of a number of professional courses addressed the exact problem they were getting from their briefers: Good information, not enough analysis. So, the powers that be re-focused the school and made it analytically focused. The analysis filed has changed dramatically in the past few years and as a result is taking more of a pattern recognition and pre-emptive strategies approach.

So I don’t think we should be worried at all, or need to hone our memories in the abstract for bits of data. In fact, with the error prone nature of our mental hardware I would suggest that it does a disservice to factual recall in trying to do so. Instead, we can “outsource” the storage to a more accurate replication medium (like a reliable computerized bookmark favorites) and apply reasoning to this complex data in its unchanged form while re-programming our internal brain data storage to be, rather than the data itself, what database what we want to know is under (The “folder” it’s in rather than the contents of the “file”) .

The question becomes how do we make the card-catalog of our mind faster and more efficient?

AndrewKemendo Jun 20, 2009, 11:29 AM
0 points
on: Cascio in The Atlantic, more on cognitive enhancement as existential risk mitigation
This comment seems to miss the idea:

What happens if such a complex system collapses? Disaster, of course. But don’t forget that we already depend upon enormously complex systems that we no longer even think of as technological. Urbanization, agriculture, and trade were at one time huge innovations. Their collapse (and all of them are now at risk, in different ways, as we have seen in recent months) would be an even greater catastrophe than the collapse of our growing webs of interconnected intelligence.

If in fact the future is what the rest of the article envisions, a world of accurate measures and prudent predictions, then the possibilities for collapse will become less and less.

Making the case that such largess will of course lead to the linear probability of increase in damage that would result in collapse, ignores in large part, if not the majority of the science behind cognitive development and AI science—risk mitigation and error elimination.

AndrewKemendo Jun 16, 2009, 2:44 AM
7 points
in reply to: hrishimittal’s comment on: Readiness Heuristics
The Yudkowski worship is getting pretty thick around here:

http://lesswrong.com/lw/4g/eliezer_yudkowsky_facts/

Lets not turn this into a fandom

AndrewKemendo 8 Jun 2009 23:28 UTC
2 points
in reply to: bill’s comment on: Post Your Utility Function
I think this greatly oversimplifies the issue. Whatever my response to the query is, it is only an estimation as to my preferences. It also assumes that my predicted risk will, upon the enactment of an actual deal, stay the same; if only for the life of the deal.

A model like this, even if correct for right now, could be significantly different tomorrow or the next day. It could be argued that some risk measurements do not change at intervals so fast as would technically prohibit recalculation. Giving a fixed metric puts absolutes on behaviors which are not fixed, or which unpredictably change. Today, because I have lots of money in my account, I might agree to your deal. Tomorrow I may not. This is what I mean by intuitions—I may think I want the deal but I may in reality be significantly underestimating the chance of −50 or any other number of factors which may skew my perception.

I know of quite a few examples of people getting stuck in high load mutual funds or other investments because their risk preferences significantly changed over a much shorter time period than they expected because they really didn’t want to take that much risk in their portfolio but could not cognitively comprehend the probability as most people cannot.

This in no way advocates going further to correcting for these mistakes after the fact—however the tendencies for economists and policy makers is to suggest modeling such as this. In fact most consequentialists make the case that modeling this way is accurate however I have yet to see a true epistemic study of a model which reliably demonstrates accurate “utility” or valuation. The closest to accurate models I have seen take stated and reveled preferences together and work towards a micro estimation which still has moderate error variability where not observed (http://ideas.repec.org/a/wly/hlthec/v13y2004i6p563-573.html). Even with observed behavior applied it is still terribly difficult and unreliable to apply broadly—even to an individual.

AndrewKemendo 7 Jun 2009 4:33 UTC
0 points
in reply to: conchis’s comment on: Post Your Utility Function

It’s only if they’re also risk neutral with respect to experienced utility

I am curious how this would look in terms of decisions under experience. Does this imply that they are expecting to change their risk assessment once they are experienced?

AndrewKemendo 7 Jun 2009 4:27 UTC
0 points
in reply to: bill’s comment on: Post Your Utility Function
Unfortunately the better parts of my post were lost—or rather more of the main point.

I posit that the utility valuation is an impossibility currently. I was not really challenging whether your function was exponential or logarithmic—but questioning how you came to the conclusion; how you decide, for instance where exactly the function changes especially having not experienced the second state. The “logarithmic” point I was making was designed to demonstrate that true utility may differ significantly from expected utility once you are actually at point 2 and thus may not be truly representative.

Mainly I am curious as to what value you place on “intuition” and why.

Testing it against my intuitions

AndrewKemendo 6 Jun 2009 22:29 UTC
0 points
in reply to: bill’s comment on: Post Your Utility Function
It makes sense however you mention that you test it against your intuitions. My first reaction would be to say that this is introducing a biased variable which is not based on a reasonable calculation.

That may not be the case as you may have done so many complicated calculations such that your unconscious “intuitions” may give your conscious the right answer. However from the millionaires biographies I have read and rich people I have talked to a better representation of money and utility according to them is logarithmic rather than exponential. This would indicate to me that the relationship between utility and money would be counter-intuitive for those who have not experienced those levels which are being compared.

I have not had the fortune to experience anything more than a 5 figure income so I cannot reasonably say how my preferences would be modeled. I can reasonably believe that I would be better off at 500K than 50K through simple comparison of lifestyle between myself and a millionaire. I cannot make an accurate enough estimation of my utility and as a result I would not be prepared to make a estimation of what model would best represent it because the probability of that being accurate is likely the same as coin flipping.

Ed: I had a much better written post but an errant click lost the whole thing—time didn’t allow the repetition of the better post.

AndrewKemendo 6 Jun 2009 18:40 UTC
1 point
in reply to: bill’s comment on: Post Your Utility Function
How have you come to these conclusions?

For example:

The reason is that, for me, changing my wealth by a relatively small amount won’t radically change my risk preference, and that implies an exponential curve

Is that because there have been points in time when you have made 200K and 400K respectively and found that your preferences didn’t change much. Or is that simply expected utility?