Free Stanford AI Course

This October, anyone can take an Introduction to Artificial Intelligence class, taught by professors at Stanford, for free.

This is great. But it gets better. In order to expand the scope of the class from the 200 people they’ve been teaching in person, the instructors will be using AI software to grade homework, aggregate discussion questions, and generally mediate interactions with students. Why bother? Because, to date, over 100,000 people have signed up for this course, and the enrollment is showing no signs of slowing.

The use of the software to scale allows students to get feedback on their individual homework assignments and quizzes, to interact with the instructors, and to get a ranking in the course compared to both other online attendees and the students enrolled at Stanford — feedback that would be utterly impossible to provide to that number of people if the instructors didn’t have the help of AI. It will be fascinating to see how the concepts taught in the class are used to administer it.

I’ve been intrigued by AI ever since reading Gödel, Escher, Bach way back when I was a teenager (and before I was really equipped to follow all of it). More recently, I’ve been increasingly interested in robotics and the applications thereof, which rely pretty heavily on some of the AI concepts that are in the syllabus for this class. And, of course, I have an enduring interest in games, which are probably where AI is used most often in modern computing.

So am I signing up? You betcha. And I hope some of my nerd friends will too, so that we can compare notes along the way.

This is one of the reasons I love living in the future: we have access to information and learning that is unparalleled in human history, opportunities to sit at the feet of experts that we could only have dreamed of even a decade or two ago. And wonderfully, access to an amazing education is increasingly being divorced from access to money, creating remarkable opportunities for people who are ready to work at their own learning, regardless of their backgrounds. I fully expect to be bested in the course rankings by smart 14 year olds in India and China, and will be excited to see it happen.

I’ll post some updates and reflections along the way, and possibly homework assignments too if they turn out to be interesting. This should be a fun ride.

(Thanks to Singularity Hub for the tip-off.)

Mad Science: Jam Jar Jet

This past Friday, Jason Young and I got together for a visit. Since Baylor Sing had just wrapped up for him, and I was starting 3 days of unemployment, we decided to drown our sorrows with alcohol.

But, being the mildly destructive nerds that we are, we decided it would be more fun to set it on fire than to drink it. So we printed our William Gurstelle’s excellent article from Make Magazine #5, gathered the necessary supplies, and got to it. Here are the results:

SuperNerd!

Last night I posted to Twitter:

Watching a Nova show on string theory. (It was a requirement to maintain my supernerd certification this month.)

This morning, my friend Jeff presented me this, created in cooperation with his wife Fazia:

SuperNerd Certificate

This of course immediately became one of my new favorite things. Be sure to appreciate the seal and the signatures!

Radio Silence

Hey, Mouseketeers! Sorry for not writing more lately. While school’s in session, our weeks tend to be busy but homogeneous, and our weekends totally unpredictable, neither of which is much good for the discipline of getting things down on paper (or electrons). Here’s the latest:

  • After returning from the conference in Springfield (where I had a lovely time, thanks for asking), I took all the kids off to the Texas Renaissance Festival with my trusty brother, my cousin and her husband in tow. This worked out well, as Liam was feeling rather sick in the morning, and the other kids were able to run around with the adults while Liam and I sat quietly under trees and watched people in chain mail walk past. (“Holy cow, did you see that 10 foot tall dragon?” “Really? You dressed as the Joker to go to a RenFest?” “Please, ma’am, you’re not really equipped to be wearing that!”) About midway through the afternoon, Liam’s belly stabilized, so we were able to run around to see the Mud Show, the falconry, jousting and fireworks. Great fun! If you’d care to follow along in the photo storybook, you can do so here. The Festival runs for several more weeks, so if you’re in the area, I highly recommend a visit. Tell them Sean sent you! (They’ll act like they don’t know what you’re talking about, but don’t worry, that’s just part of the security.)
  • Last weekend was Maker Faire, which is kind of like a giant support group for people like Jason Young and me who like to build large, potentially lethal do-it-yourself projects. The three younger kids and I met up with Jason and Erin, his (far) better half, for a great day viewing art cars, making shrinky dinks out of recycled plastic, riding bicycle animals, watching robots fight to the death, learning to weave on a loom, meeting the Eepybird guys, dodging huge gouts of flame shooting from fire plugs, seeing rockets launched, crocheting handbags from used plastic grocery sacks, and more. Photos are here, and I’m editing down some of the video I got for a highlight reel, to be posted soon.
  • This morning I went and laid down some pennywhistle tracks for an album some friends are putting together. Fun to get to take that into the studio, though I always forget how exacting recording can be, even in short stints. John, the audio engineer, had 7 mics on the whistle, so I’m interested to see what sort of tone he manages to extract from that daunting array. Between that at the Thursday night Irish Sessions, I’ve been enjoying keeping a foot in the musical world even during The Patio Boys’ recent hiatus.

Other than that, it’s school, school, school! The kids report cards are trickling in for the first six weeks, and they all seem to be doing fine. Kathy’s been kicking butt and taking names at the University — the former figuratively, the latter literally, as she’s heading up a couple of student organizations for which she’s been assembling T-Shirt orders. And I’m still plowing away at the day job, leading a rag-tag fleet on a lonely quest across the galaxy to find Earth, our long-lost home planet.

Doggoneit. That’s not me.

Robocars!

I’ve been intrigued by Personal Rapid Transit since 2002. I’ve gone so far as to do some software modeling and simulation of a PRT system for San Marcos (with a station right by my house, natch), and those of you unfortunate enough to live near me have probably been regaled by my excited blathering on the subject from time to time. (“Good goshamighty, why won’t he be quiet? Shut up! SHUT UP! He won’t stop. Ok, I’m going to my happy place now. Thinking about unicorns and cheesecake. Mmmmmm…cheesecake.”)

But as of today, I’m no longer a fan.

The reason? Not any particular deficiency on PRT’s part. It’s still quicker, more efficient, and generally much more desirable than private cars or other public transit options.

But today, I discovered Robocars.

What are Robocars? Cars that can drive themselves. DARPA’s Grand Challenge in 2004 and 2005 showed that autonomous vehicles were a technological possibility. The 2007 Urban Challenge demonstrated that they could function on normal roadways, dealing with traffic laws, other vehicles, and pedestrians along the way.

Once you have vehicles that can drive themselves, a fleet of them (run by, say, a taxi company) offers nearly all of PRT’s advantages without its drawbacks. Consider these improvements on the PRT concept:

  • Robocars operate on existing roadways. There’s no need for a dedicated infrastructure like PRT’s.
  • If a Robocar breaks down, it can easily be moved out of the way like a car. It wouldn’t block an entire monorail line like a failed PRT tram would.
  • Implementation of Robocars would not need a huge initial investment from a central government, but can be introduced one at a time, just like automobiles. It can therefore happen as gradually or rapidly as the market dictates.
  • By communicating with traffic lights and ensuring optimum travel speeds, Robocars could be nearly as efficient as a PRT system.
  • One can have personal ownership of a Robocar if one wishes.
  • Robocars remain useful as you move out of urban areas.

Robocars, like PRT, offer even more advantages over traditional cars and transit systems: fuel efficiency, convenience, safety, flexibility, and the ability to both enjoy a private space and to do other things while in transit.

See Brad Templeton’s Introduction to the idea for an overview of what it’s all about, or for a glimpse into the possibilities, read his fictional account of what a week with Robocars might look like.

Sci-Fi Future: Bioengineering

On one of our recent dates, Kathy and I had stopped by the local pet store to browse around a bit. While passing by the fish, I noticed tank full of fish that were even more brightly colored than the usual tropicals. When asked, a salesperson explained to me that they were GloFish: zebra fish that had been genetically engineered to include a fluorescing protein created by a jellyfish gene. Originally created with an eye toward detecting toxic chemical spills, they are even more eye-catching than the photos show.

The next day, I was listening to an episode of WNYC’s excellent Radio Lab program where they discussed some young bioengineers who got tired of having to smell E. Coli, which is notoriously poopie-scented, all day in their lab. They began by introducing wintergreen genes, and soon had minty-fresh E. Coli in their lab. They then went a step further by having the bacteria start producing a banana smell when full grown, so that the scientists could tell if a culture was ready for experimentation with the merest whiff.

And of course, we’ve had genetically modified foods on our supermarket shelves since the early 1990s. Various GM varieties are more disease and pest-resistant than their unmodified counterparts, have higher yields, last longer without added preservatives, and have their vitamin content boosted.

So, in many ways, it seems like we’re at the dawn of a golden age of bioengineering. We’re able to improve on naturally grown foods, we can engineer unpleasant characteristics out of experimental organisms, and we can even tailor our pets to make them more interesting and fun. What’s not to like?

Quite a bit, as it turns out. Lots of people have concerns about bioengineering, and wonder if it may be a Pandora’s Box we might wish closed again once we have pried out its secrets. A few points to consider:

GloFish are patented just like a mechanical invention would be. From their FAQ:

Because fluorescent fish are unique, their sale is covered by a substantial number of patents and pending patent applications. The providers of GloFish® fluorescent fish, 5-D Tropical and Segrest Farms, are the only distributors that have the necessary licenses to produce and market fluorescent fish within the United States. The production of fluorescent fish by any other party, or the sale of any fluorescent fish not originally distributed by 5-D Tropical or Segrest Farms, is strictly prohibited.

The fact that this patent was granted to cover not just a mechanical device or invention, but a form of life, seems like a pretty big leap. (And allows them to charge an order of magnitude more for these GloFish than for their unmodified brethren.) What do we do with patent and copyright law as we plow into this new area of human endeavor? Consider, for example, an excerpt from this article:

If you could duplicate a person other than yourself, who would it be?

This is not a hypothetical question. Human cloning, may allow you to do
that, with or without the clonee’s consent. Once human cloning technology is available all you’ll need is the desired DNA, and that can be very easily obtained: It is called DNA piracy. The ease of stealing DNA for cloning purposes raises the following question: how is the law going to protect my genes and what legal remedies are afforded in such a case.

DNA Copyright Institution Inc., a privately held corporation in San
Francisco, proposes a solution. It promises copyright protection to your
genetic profile for only $1,500. The visionary DNA Copyright institute,
founded by Andre Crump, is trying to persuade celebrities to use its
services to strengthen their legal position should anyone decide to clone
them against their will.

Yep, the folks out in California are already planning for what happens if you get a strand of Cindy Crawford’s hair and decide to make your own Cindy clone using the DNA therein. More troubling, is it possible for corporations to copyright certain genetic sequences? And if so, can they then bring action for infringement against people who have those sequences in their own genome naturally? There are lots of lines to be drawn here, and it’s not always at all clear where they should be scribed.

Once we have the technology, is is OK to genetically engineer Multiple Sclerosis out of our babies? If so, what else can we change while our kids are still on the drawing board? Can we then choose eye color, hair color, and attractiveness? Could we add a few inches of height to give our kid a psychological advantage? Could we add a few more inches to give them an advantage in basketball? Should our modified basketball player be in the same league as non-modified players, or should there be a GMNBA?

And what of biodiversity? Artificial genes from GM crops can “leak” into the wild population. Even without GM, lots of farms have moved to monocultures — the planting of only the single highest-yield variety of their crop. This tendency would likely be exaggerated further if GM crops showed even better yields than their naturally occurring counterparts. This monoculture farming means that an entire crop can be wiped out by a disease to which it happens to be susceptible. Ironically, it also results in having access to less raw genetic material as the less popular strains are bred out of existence.

Finally, what happens when the bioengineers who may have more malevolent intent start fooling around with this stuff? Freeman Dyson, the futurist who conceived that trusty science fiction chestnut the Dyson Sphere, talks about children having access to home genetic engineering kits. This sounds like great fun as long as kids are just making unicorns or, as South Park would have it, a monkey with five butts.

But what happens when we start bioengineering weapons? Little Timmy could toss together a few genes from bird flu, the cold, SARS, bubonic plague, and a dash of smallpox, mix well, and viola! Instant highly-virulent superweapon! Take it further: engineer it to attack specific racial traits, and you could have a Final Solution that would cause history’s atrocities to look wan and insignificant.

It seems that we have discovered a very powerful tool here. As with all powerful tools, it enables us to accomplish amazing things that were previously impossible, but also has the potential to cause irreparable damage if used irresponsibly. Thus, while our enthusiasm here may tempt us to rush in to a Brave New GM World, I think it’s vital that we approach this new territory with caution. Pay attention to these discussions, befriend a bioethicist, and encourage our lawmakers to take these issues seriously. Our children and their unicorns are depending on us.