mgroves

Posts Tagged ‘science’

I saw this video and tweeted about it, but it’s so cool that I think it needs a blog post.

It’s a real live, operational, Turing machine (sans infinite tape, of course).

So, what’s a Turing machine?  It is a theoretical (until now) device that is capable of scanning, reading, and writing binary numbers onto a “tape” of cells (think of it as a binary array of infinite size).  Alan Turing (naturally) first described such a machine in order to provide a foundational representation of machine computing. It’s not practical, but it’s a good place to start.

Okay, it seems kinda silly, what’s the point?  The point is, well…there are a lot of points.  However, one of the most interesting things about the Turing machine is that a clear definition for “computer” can be written in terms of a Turing machine.  Formally, the Church-Turing thesis states that everything computable can be computed by some Turing machine. This can be taken one step farther: any machine that can simulate a Turing machine, is a computer. So, my laptop is a computer, my Xbox is a computer, and my calculator is a computer. But even an abacus is a computer, and even a bunch of rocks can be a computer!

Anyway, I tend to get excited about nerdy stuff like this, and I thought it was very cool that someone took down the nebulous mind’s eye view of a Turing machine and built it out with servos and a long film strip. It got me thinking about those long assignments writing out state machine diagrams and simulating Turing machine outputs on paper.

Let’s assume it’s possible to create an artificially intelligent machine, ala science fiction. What if this AI became malevolent? Why not limit its communication to a single chat with only the original programmer until safety is assured.

In that situation, it’s speculated that a sufficiently smart AI can talk its way out of its “box”, because humans are not secure.

Is this possible? Well, there’s one way to test: instead of an actual AI, use a human to simulate AI (almost like a reverse Turing Test). If a human is told not to let the AI out of the box, can a human posing as an AI convince the human to let him out of the box?

The opposing viewpoint is so: “There is no chance I could be persuaded to let the AI out. No matter what it says, I can always just say no. I can’t imagine anything that even a transhuman could say to me which would change that.”

This experiment has been run on two occasions, both resulting in the human letting the AI out of the box:

• First time
• Second time

However, this test is very suspicious for a number of reasons. First, Eliezer Yudkowsky, the guy who proposed this test in the first place, is the one simulating the AI in both cases. Second, the actual chat transcripts between the “transhuman AI” and the human are not publically available (as far as I know).

But let’s assume for a minute that there are no shenanigans going on. A human knowingly participating in this experiment, knowing the objective, knowing the stakes, and yet letting the AI out anyway is a frightening prospect. What does this say about humans-as-weakest-link in IT security, let alone the nature of humanity?

More importantly, what could an AI, restricted solely to a single one-on-one chat room with you, possibly say to convince you to let him out of his box?

IDEO is a design consulting firm. They have designed many of the things you use today. Probably some of the things you are using right now.

They’ve designed the first Apple mouse, the Palm V, and the Oral-B toothbrush gripper. They also designed a revolutionary shopping cart in 5 days…

They are no doubt the best design firm in the world, receiving more Industrial Design Excellence Awards than any other firm.

They are also total hippies:

• Idealized egalitarian meritocratic work environment? Check.
• Sticking it to “corporate America”? Check.
• “The basket is tyranny”? Check (19:15).
• Patrons of Whole Foods? Check.
• Can you naturally concatenate every sentence with “maaaaan”? Check.

Also, what’s the deal with Ted Koppel? His face looks like a hastily assembled Mr. Potatohead.

I’m back, baby! Back to blogging at mgroves.com again, that is. Just a reminder, if you want to know what I’ve been up to all this time, check out MatthewKGroves.com.

Now, on with the show!

What’s new with mesothelioma these days? Here are some links to previous mesothelioma updates:

• Mesothelioma
• The Return of Mesothelioma
• What is mesothelioma?

Well, according to a “brain expert”, the effects of mobile phones are worse than asbestos (and I assume the resulting mesothelioma) and smoking. How sensational!

Does this mean that smoking isn’t nearly as bad as those infuriating Truth commercials make it out to be, or does it mean that using cell phones will kill us all?

Discuss whilst I light up some cigars today to celebrate the birth of my son.

Much hay has been made about the CFL (compact fluorescent lamp) “bulb”, which includes a lot of “green” talk and a lot of talk about saving money on your electricity bill.

I won’t spend any time on the “green” arguments, because doing so begs a lot of questions, but what are the economic benefits of using CFLs, if any? I did an analysis like this for Hybrid cars, especially the “hacked” hybrid, some time ago. Now might be a good time to revisit that analysis to discover that, no, hybrids still don’t make sense (in general).

Let’s start with some facts and try to do an apple-to-apple comparison. The 13 watt CFL is the “equivalent” in terms of light of a 60 watt traditional bulb. Since the 60 watt bulb is the most common, I will use these bulbs as comparision.

Additionally, I will use the cheapest price of these bulbs: 60 watt bulbs you can get 4/$1.00 ($0.25), 13 watt CFLs are 12/$19.76 ($1.65). CFLs have longer lives too: I’m using 750 hours of life for 60 watt bulbs and 6000 hours of life for CFLs. The only other data we need is the number of hours per day that the bulb will be on, and the cost of electricity. I’m going to assume 5 hours a day, because I’m picturing the lights in my living room. I found the cost of electricity by kWh here; for Ohio it’s 9.51 cents/kWh.

I will give you the spreadsheet if you want to tweak these values (if you live outside Ohio, can get cheaper/longer lasting bulbs, whatever), but just go with me for now.

Here’s the results:

This means that if you plan to use a light source for more than 62 days at 5 hours a day, you are better off using a CFL.

But not every light in your house is on that long. So at what threshold should you use a CFL? Assuming a 1-year planning horizon, here is the answer:

This means: if you are using any light source in your house for more than 0.4 hours per day, every day for a year, you should probably use a CFL.

The amount of savings here isn’t drastic: at 5 hours a day, you will save $7.51 a year. At 0.5 hours a day, you’ll save $0.17 a year. Any light that’s on for less than that (every day for a year) will lose you money.

In the long run, it’s always better to us CFLs, as they will always outpace a light bulb, even if it takes a few years. So, if you are going to live where you are for a while, or you are comfortable with taking all your CFLs with you when you leave, it makes sense to get CFLs for everything. Otherwise, get CFLs for high usage areas, and use old bulbs for low usage areas.

Some additional comments: my current supply of standard bulbs (in use and in the package) does not enter into this analysis, because I consider them a sunk cost. Additionally, there might be some macroeconomic benefit in the long run by buying CFLs because it will allow companies to take advantage of economies of scale. That is beyond the scope of this analysis.

One press of a button and you can end your life with a swift injection of potassium chloride. That is the boast of Roger Kusch, once one of Germany’s most promising conservative politicians and now the improbable promoter of a mercy-killing machine.

Bender: “Listen, buddy, I’m in a hurry here. Let’s try for a twofer. Hehe.”
Suicide Booth: “Please select mode of death. Quick and painless or slow and horrible.”
Fry: “Yeah, I’d like to place a collect call?”
Suicide Booth: “You have selected slow and horrible.”
Bender: “Great choice!”

I wonder if this relates to the deliberate, hesitant pessimism in Germany.

Speedlinking is yet another tool of the lazy blogger. Basically I smash together a bunch of interesting links that I collect every so often, and write a sentence or two about them.

This week’s Speedlinking is being brought to you courtesy of the useful Instapaper website, which I’ve been using to temporarily bookmark the below links.

• Hail, Xenu! Use it to find broken links on your site.
• Drink tea? Here are some useful things to do with those used teabags.
• The Wall Street Journal reports on the many links between Saddam Hussein’s government and terrorism that are discussed in a recently released Pentagon paper. The most interesting thing to me is that terrorism was really Saddam’s only “tool”, because of the sanctions and no fly zones and what not. Unintended consequences and all that.
• The number 6174 is actually pretty awesome.
• If you tinker around with the URL, you can actually increase the quality of playback on YouTube videos. I guess YouTube went ahead and made this an account option too, so you can always default to “HD” if you want.
• ha.ckers.org looks at an interesting method of spamming: exploiting the ability to change email addresses without confirmation.

I just had my wisdom teeth removed today.

I must say that I recommend nitrous to anyone considering having this procedure. It’s much like being drunk, instantly, and without a hangover. Very nice.

The top two represent my former wisdom teeth.

So far, no real pain. My lower lip and chin feels like a dead slab of meat, which makes drinking and eating hilarious. Bleeding is slowing down.

The biggest complaint I have is the whole ice pack thing: 20 minutes on, 20 minutes off, FOR 24 HOURS. Crap.

Other than that, I haven’t touched the pain pills. Yet.

I took a full year’s worth of Google data and smashed it up with data about mgroves.com blog posts to see what (if anything) I could learn.

I ran a regression, and it’s telling me that my humor is not appreciated.

First, let me very briefly explain regression. It’s a statistical modeling technique in which a dependent variable (daily adsense revenue in my case) is analyzed as a function of one or more independent variables (in my case this includes things like common tags, number of comments, and clicks). The idea is that the independent variables explain the variance in the dependent variable. This isn’t a great explanation, but it will do for now.

So what I did was I matched up Google AdSense data (date, revenue, clicks) by day to my blog post data (date, keywords, number of comments) per day. So each row counts as a day, not a blog post.

I then ran a regression for 2007. Here are the results.

Some quick interpretation:

• Adjusted R Square: this tells me how much of the variance of the revenue can be explained with this model. In this model, it’s around 54%. Pretty darn good!
• Significance F: this number tells me the likelihood of this model not being statistically relevant. Very low numbers are good.
• Coefficients: This tells me how much each variable affects revenue. So, looking at “wii”, for instance, tells me that a post with a tag of “wii” affects revenue by +28 cents.
• P-value: This column is similar to Significance F. It tells me the likelihood that using this coefficient will give me an inaccurate result. The lower the better, but I chose a cut-off point of around 0.10. 0.05 is a popular cut-off point too. Any P-value higher than the cut-off and the variable really doesn’t tell me anything one way or another.
• Confidence interval (Lower & Upper 95%): The coefficients give me a single point estimate, but this range indicates that it’s 95% likely that the actual coefficient is in the range. If a range straddles 0, it means that this coefficient isn’t useful even in a basic positive or negative sense.

So based on that, what does this regression tell me?

1. “Wii” posts really bring in the money. 28 cents each!
2. My jokes aren’t funny. This data went out of its way to tell me that “humor” posts cost me 12 cents of adsense revenue each. Good P-value and a confidence interval that’s all negative. Thanks for the self-esteem boost, data!
3. AdSense clicks are correlated with revenue. Duh.
4. I ran some other models with less variables, and they indicated that “politics” and “programming” might be worth some money (around 14 and 20 cents each). The model shown above doesn’t indicate anything useful about either.

There are some flaws with this model. The biggest one is that the data is lagged. While a blog post might go up on Tuesday, no one might actually click ads or post comments for a day or two (or even much, much later for that matter). I might try lagging the dates by 1 day and 2 days to see what a regression says then, or maybe matching up comments by the day they are posted (instead of the blog post they are attached to).

Ah, Chrono Trigger. One of the finest RPG’s of all time, and perhaps my favorite.

The story is this: Crono and friends accidentally discover time portal(s) and get flung into a post-apocalyptic future caused by Lavos, a giant parasite created and summoned by an evil mystic named Magus.

When Crono and his party are first flung into the future, they enter one of the ravaged cities known as “Arris Dome”. They find a few disheveled survivors who are able to (painfully) survive without food through the use of an enertron. Supposedly there is a huge store of food in the lower levels of the Arris Dome, but it’s being guarded by robots, and the last man who tried to get past them never returned.

Crono and his team are able to defeat the robots, only to find that the refrigeration has failed and all the food has spoiled.

They do find the body of the last man who tried to retrieve the food. He is clutching seeds that could be used to grow food, thus giving the survivors a glimmer of hope for their devastated world.

But did you know that a real Arris Dome has been created right here on Earth in an Arctic mountain near the North Pole?

That’s right. When the Day of Lavos comes to Earth, we’ll have a sample of seeds from more than 1400 seed banks from which to help us survive.

But wait, what if the refrigeration fails?

“The seeds of wheat, maize, oats and other crops will be stored at a constant temperature of minus 18 degrees Celsius, and even if the freezer system fails the permafrost will ensure that temperatures never rise above 3.5 degrees Celsius below freezing.”

Phew, good. Now all we have to worry about is the guardian robots.

Bring it on, Lavos, you overgrown porcupine!