Sunday, July 31, 2005

Technology Review against SENS

Jason Pontin, the editor of Technology Review has issued a challenge:
submit an intellectually serious argument that SENS is so wrong that it is unworthy of learned debate, and you will be paid $20,000 if it convinces independent referees.
SENS (Strategies for Engineered Negligible Senescence) is Aubrey de Grey's program to . de Grey has put a fair amount of intellectual effort into this fight, and has produced an argument that is convincing to some, but apparently repugnant to others. (See the article in TR, and an editorial by Pontin in which he made it clear that he found de Grey's ideas distasteful, but didn't have anything to say about the science.

The statement above is curious for a scientific inquiry. Pontin isn't asking whether SENS is possible, or might be debugged so as to reach de Grey's goals, he wants someone to show that no one should listen to de Grey or try to convince others that he's wrong. It seems like Pontin has proven the point he's trying to defeat. If it takes $20,000 to convince gerontologists (or if that amount is insufficient) to come up with an intellectually serious argument, then isn't the question worthy of serious debate?

My impression is that serious physicists stay away from claims about perpetual motion machines because they could spend all their time debunking them with no gain. And it would be a pure loss to a particular physicist if he came up against a kook who had built a device that the physicist couldn't figure out in a reasonable amount of time.

But de Grey's proposition is different. He's written serious articles in serious journals, co-authored by serious experts arguing that various aspects of his program make good engineering sense, comport with the standard biological models, and would move us toward longer healthy lives. There aren't an endless supply of other crackpots that the gerontologists would have to debunk if they accepted de Grey's challenge.

Pontin states his rules clearly and simply. #2 is

The purpose of the Challenge is to establish whether SENS is worthy of serious consideration. Submissions are sought that attempt to demonstrate that it is not.
People of the opposite view need not apply. It sure doesn't look like the purpose of the Challenge is to establish whether SENS is worthy of serious consideration.

Pontin's reply at FightAging.org is more ambivalent.

  • Indefinite life might be good for me, and I might wish it for those that I love, but an entire world of superagenarians might be a bad thing.
  • I am not sure significant life extension is possible [...]
  • But if SENS is reasonable, it's obviously very important news, and worthy of serious attention[...]

I got my first pointer into the discussion from the extropians chat list, but there are enough pointers here to connect you to many threads of discussion all over the net.

Wednesday, July 27, 2005

Tools I use: Ant

I've been using , a build tool for Java. It works pretty well, and overall I'm glad I'm using it. Zocalo builds cleanly on my desktop Mac, my PowerBook, and the linux server at CommerceNet. Any of the platforms will compile, test, and build javadoc or tar files, even when I couldn't find a copy of javadoc without a lot of effort. Ant is much more flexible and powerful than the special purpose shell scripts we used at Agorics. We had all previously used Make, so we knew that that style of approach had its advantages, but Make apparently doesn't really work with complex Java projects, so it's nice that Ant has been built to fill that gap.

But I just don't understand the fascination with XML. XML is a really lousy language for people to interact with directly. I could understand it if people wanted to use XML as an interchange language between programs, but it doesn't seem to have any advantages over s-expressions. The extra syntax just leads to extra possibilities for mistakes, without any value when edited by hand. Shouldn't Ant ship with an editor that lets you fill out a form to create your ant description? Why do we have to hand edit the XML?

Anyway, I also wanted to point out the solution to one of my problems. It took a while to find the right way to invoke . There's apparently a common problem caused because the ant task for JUnit relies on an optional package that isn't in the standard classpath. Rather than mucking up your classpath for your project with ant's optional tasks, you can just write a recursive call on ant that sets the appropriate classpath and then invokes the simple JUnit task. The details are given on the Open Source Lab's wiki. The recursive call is simple:

<target name="test" description="Run JUnit Via Exec">
   <exec executable="ant">
       <env key="CLASSPATH" path="${basedir}/lib/junit.jar"/>
       <arg line="testTask"/>
   </exec>
</target>

Once you've done that, the JUnit task works correctly, so you can specify which tasks to run using a fileset pattern, rather than enumerating all the test classes as you'd have to do if you tried to work around the classpath problem by execing JUnit directly.

I added one thing to Open Source Lab's suggestion: I wanted a short listing of failures printed out, so I grepped JUnit's output files. All I had to do was add these lines before /target:

<exec executable="egrep">
   <arg line="-r" />
   <arg line="'(Failures: [^0]|Errors: [^0])'" />
   <arg line="TEST" />
</exec>

Thursday, July 21, 2005

Tools I Use: realage.com

I'm going to occasionally write about tools I use that do their job well. These can range from software to garden tools to tips and tricks that I think will be new to other people. Most often, they'll be software of one kind or another. Today's tool falls in the category of and .

I have several sources that I rely on for health advice: Janet (my S.O.), Health Magazine, and realage.com are the primary ones. Health Magazine used to be aimed at a wide audience, but now is more focused on women. It still has a fairly high standard for articles on health-related issues. All I have to do is skip over the beauty tips, and the rest is fairly useful. (Their main sections are Body, Mind, Food, Beauty, and Fitness.) There's more relationship advice than I need, but it's about getting along with your boss as often as about getting along with your mate, and they don't have silly articles about how to attract the perfect man, so I don't object to it. As I said, their standards are fairly high in terms of insisting that their writers cover subjects where there's research behind the advice. They don't alays give sources, but often do refer to the scientists behind the findings, or even interview them.

The other source I go to is realage.com. Their schtick is to ask you a bunch of health, lifestyle, exercise and dietary questions, and then tell you your "real age", reflecting your life expectancy as a comparative age. They tell me I'm doing many things right, so my "real age" is about 39, while I'm chronologically 46.

They also have plenty of advice about what you might change about your lifestyle to decrease your real age. The feedback and memetics on this are really good for those with a little ability for delayed gratification. The things you change about your life are reflected immediately in your expected lifespan, and apparent youth! If you backslide, you know it's increasing your real age *now*, not eventually.

I was initially attracted to this site by an article in "Health" Magazine that said that the people running the site are doing a good job of relying on and referring to the best scientific research on what changes make an actual difference in longevity. They provide references to the review papers that establish the efficacy of their advice when I've dived deeper into particular recommendations. Their policy is to only rely on results that are stable across repeated clinical trials. Their "real age" metric gives a very good feel for the effectiveness of lifestyle changes. You can easily decide whether you're willing to make different changes in your life when the benefits are expressed in a commonsense unit like an extra 6 months or 3 years of expected lifespan.

I currently take Vitamins C and E, Folic acid, and Calcium on the basis of the advice from RealAge.com. I'm slowing down to closer to the speed limit when I drive, because this makes almost a year's difference in my RealAge. I occasionally revisit the site and update my answers when something changes in my life to find out how I'm doing. I changed jobs recently, and so I'm riding my bike more often than I have in 15 years, for example.

I recommend the site.

Saturday, July 16, 2005

re: Externalities

Hal Finney posted a long note on the Exi Chat list starting out with:
Seems like a lot of the problems we face could be solved if there were some reasonable scheme to compensate people for negative externalities.
Hal went on to talk about a number of situations where there are both positive and negative externalities and wondered how one could negotiate a more economically optimal outcome. There has been a fair amount of discussion in the blogs I read about an idea called Dominant Assurance Contracts (DAC) which seem well-suited to addressing just this problem. The idea of a DAC is based on an older idea called Assurance Contracts, in which individuals sign up to fund a public good. If enough people join in all the pledges are collected; otherwise, everyone gets their money back. (This idea has actually been implemented at fundable.org.) Alex Tabarrok (who writes for the blog Marginal Revolution at marginalrevolution.com) extended this proposal several years ago. In his extension:
if the group goal is not met then everyone who offered to contribute is given their money back plus a bonus. It turns out thatit then becomes a dominant strategy to contribute and the public good is always provided!
The most recent blogging on the subject is by Mike Linksvayer. Kragen Sitaker has been pushing the idea, and wrote up a long description on his "thinking out loud" list. They've also been written up at Wikipedia.

Friday, July 15, 2005

Zocalo Released

I should mention that I open-sourced Zocalo (the Prediction Market software I'm working on at CommerceNet) earlier this week. It's now at SourceForge, with both source and binary versions. I'll be posting most of my Prediction-Market related blogs at CommerceNet's blog, but I'll try to mention them here.

Saturday, July 09, 2005

Longevity Matters

Alex Tabarrok was recently having lunch with Tyler Cowen and Bryan Caplan, and the subject of longevity came up.
At lunch with Bryan and Tyler last week the question arose as to what we would do differently if we were immortal. After a nerdy discussion to clarify what sort of immorality we were talking about; the kind where you can't be killed but can be imprisoned or the kind where you are forever young but may be hit by a truck? (it was the former) - I answered that I would travel more.
They chose the wrong question to answer. Both Alex and Tyler are familiar with Cryonics, and Bryan appears to have discussed health and longevity with Robin Hanson, so I'll assume he's at least passingly familiar with these ideas as well.

It's possible (some would argue likely) that many of us are approaching the latter kind of immortality, (long-term youthfulness) but the former (actual immortality) is still a science fiction dream. Alex implied that he was making choices about how to spend his time based on what he'd do if he couldn't be killed. He said that he would travel more, and at the end of the post he said that he is going on a solo trek to Machu Picchu.

Maybe he would have come up with the same answer if they had addressed the right question. I hope so, since a solo trek in the Andes might be just the kind of risky activity that you'd get a different answer for. Travel more certainly, but a high chance at prolonged health wouldn't convince me to add significantly to my risk level. If Alex is an experienced high-altitude hiker this may not be that high a risk. But for most people, I'd guess that solo hiking in the Andes would cut into your chances of enjoying that long-term youthfulness that makes you think about travelling more widely.

Thursday, July 07, 2005

Why PanCrit.org?

I grabbed pancrit.org several years ago, when registering good names was all the rage. Since then I've used it occasionally, but used mydruthers.com much more often. Pan Critical Rationalism is the epistemology described by W. W. Bartley III, in his book The Retreat to Commitment. I wrote about Bartley's epistemology on my Books page when I originally set up a web site way back when. Having a good epistemology provides the foundation for thinking about anything. And I like the theme of "criticize everything" as a foundation for a blog. Let's start there.