As we have moved through the 20th century and into the 21st, and computer processing power has increased exponentially, popular culture has been progressively gripped by the what-if scenario of computing machines becoming sentient. Fictional examples run from saviours to annihilators, but all characterizations hinge on one assumption: that computing machines can become sophisticated enough to accurately replicate the processes of the human brain – and therefore spring to ineffable internal life. Some thinkers, including Stephen Hawking: see the transition happening so soon that they have started publicly warning us away from developing this “Strong A.I.”
But over at Psychology Today, cognitive neuroscientist Bobby Azarian argues that strong A.I. — that is, the kind that we puny humans need to worry about taking over the world and wiping us out — is a non-starter; what he describes as a “myth.” And it has everything to do with what plagued Jill Watson in the tale we related last week: the gulf between the ability to replicate, which computers can do quite well, and the ability to understand.
This gulf has been proposed to be technical in origin. Due to the binary decision making process that is the deepest foundational building block of even the most sophisticated level of computing:
“[A] strict symbol-processing machine can never be a symbol-understanding machine. […]
[Physicist and public intellectual Richard] Feynman described the computer as ‘A glorified, high-class, very fast but stupid filing system,’ managed by an infinitely stupid file clerk (the central processing unit) who blindly follows instructions (the software program). Here the clerk has no concept of anything—not even single letters or numbers. In a famous lecture on computer heuristics, Feynman expressed his grave doubts regarding the possibility of truly intelligent machines, stating that, ‘Nobody knows what we do or how to define a series of steps which correspond to something abstract like thinking.’
While a powerful computer can potentially replicate the physical processes of the human brain exactly in a virtual space, there is always something missing from it to prevent it from coming into full, self-aware life of its own. Scientists speculate that this missing part may be electrochemical, unique to the organic machines that we call out brains. Which means there is hope for humanity’s survival yet!
It’s usually a shock to the system when someone transitions to a higher level of education. With a more competitive atmosphere and higher student-faculty ratio, it’s easy for basic questions to go unanswered, and connections to professors — and subjects — to be lost.
But Ashok Goel, instructor of the Knowledge Based Artificial Intelligence masters-level course at Georgia Tech, refused to let that happen to his students. And, at the same time, he gave them an object lesson about the capabilities of the sorts of artificial intelligence they were studying.
Goel and his eight teaching assistants created a program to answer the most common questions students posted on the KBAI class forum, and disguised the program with the human name Jill Watson. Jill, presented to the class as the ninth TA, interacted with the students online for the rest of the semester. While a few students suspected something, no one fully realized she was a bot until Goel and team announced it at the end of term!
They had to start her off slowly, as she was excellent at recognizing content, but terrible at context:
“‘Initially her answers weren’t good enough because she would get stuck on keywords,’ said Lalith Polepeddi, one of the graduate students who co-developed the virtual TA. ‘For example, a student asked about organizing a meet-up to go over video lessons with others, and Jill gave an answer referencing a textbook that could supplement the video lessons — same keywords — but different context. […]’
After some tinkering by the research team, Jill found her groove and soon was answering questions with 97 percent certainty. When she did, the human TAs would upload her responses to the students. By the end of March, Jill didn’t need any assistance: She wrote the class directly if she was 97 percent positive her answer was correct.”
In April, when Goel and the other TAs revealed all, the students loved it — subsequently forming alumni groups to keep studying Jill, and to attempt to recreate her. Thanks to Jill’s success, the KBAI class will have another robo-TA next year (under a different name, of course!), to help instruct the students in more ways than one.
I love trying out new recipes: as long as I have a sharp knife and a good source for unusual produce, I fell well equipped to give almost anything a shot! But the requirements for this new sandwich gave me pause, as I’d need more than my kitchen — I’d need a lab.
The “sandwich” in question is a supercapacitor that is designed to power ingestible electronics, and is therefore made out of food. Yes, literal food! Ingredients include cheese, nori, gold leaf, and Gatorade. From Smithsonian Magazine:
“The steps for making the supercapacitors—the recipe, if you will—go like this: researchers mix a bit of egg white with carbon pellets (activated carbon, sometimes called ‘activated charcoal,’ is used in some digestive medicines), then add water and more egg white. They apply the mixture to a bit of edible gold foil. They then layer together a slice of cheese and a sheet of gelatin with the egg-and carbon-covered gold foil. On top of that they add a square of dried seaweed, the type used to roll sushi, which has been soaked with drops of energy drink. They stack more of the same materials together, and seal them in a sealing machine.”
The ingredients in this supercapacitor, when combined as above, can store and conduct electricity just as well as traditional components made out of indigestible graphene or polymers. But they have the added advantage of not needing to be passed from a subject’s system — they are just plain eaten, and experience the same fate as a regular mouthful of food. The only downside is that they need a bit more development in order to make them smaller: current prototypes are about the size of a ketchup packet, and in order to work, the supercapacitors have to be swallowed whole.
What I find really fascinating about this invention is that it foregrounds the mechanical nature of the human body, and how fuel for us can easily double as fuel for, say, a small camera taking pictures of your stomach lining. I look forward to the day when a cheesy, seaweed-y snack can do more for me than entertain my taste buds — it can help monitor the state of my insides!
Now that Victoria Day, the traditional date in Southern Ontario that marks the end of the chances of frost, has passed, I’m allowing myself to get attached to the plants in my garden. And boy, are they (and their friends in the woods around us) really starting to do their thing!
While the happy spring plants that appear almost overnight can seem to be magic to winter-weary eyes, we all know they’re actually the result of a no less stunning scientific process: photosynthesis. Nature has perfected this energy transference system, and researchers have been striving to replicate it for our own purposes. Until now, we have had less-than-efficient results.
But a team out of Harvard University and Medical School has gotten the closest to true artificial photosynthesis yet, publishing their results in Science. They call their innovation the “bionic leaf,” and it uses solar energy to combine the components of split water molecules, and hydrogen-eating bacteria, into the kind of fuel useful to humans.
The design is an improvement on the previous version, which used the process to create isopropanol. But it did so at the expense of the bacteria that were central to the operation, when they were attacked by a byproduct of the catalyst used to produce their own dang hydrogen. (It’s not easy being bacteria!) The high voltages required to circumvent this problem rendered the process too inefficient for widespread use.
The new version, with its non-bacteria-toxic cobalt-phosphorus alloy catalyst, allows for lower voltages, increasing efficiency to a stunning 10%. (The most eager plants out there hit a rate of 1% efficiency.)
“‘The beauty of biology is it’s the world’s greatest chemist — biology can do chemistry we can’t do easily,” [Prof. Pamela Silver, one of the lead authors] said. ‘In principle, we have a platform that can make any downstream carbon-based molecule. So this has the potential to be incredibly versatile.’”
Not only is this innovation just plain cool, it paves the way for fuel creation of the future: when we can finally untether ourselves from oil, and rely on the far more dependable nuclear generator in the sky for all our needs! The bionic “leaf” has it right: Mother Nature really does do it better.
As loyal readers of this newsletter know, we at DFC are advocates of making your workplace where you already are. Our “where” happens to be a cabin in eastern Ontario, but we look forward to the day when folks all over can use technological interventions to bring their workplaces to them . Until that happens, we fully recognize that most people have to bring themselves to work instead! But that means braving the dreaded commute. (*Organ riff, thunderclap*)
One strategy to help with the commute conondrum is currently being revived after making the viral rounds a couple years ago. Engineer Song Youzhou has presented a working model of his “straddling bus” concept, heretofore only existing in animated form, at the recent 19th China Beijing International High-Tech Expo. The idea behind this bus is ambitiously neat: as wide as two lanes of traffic and two storeys tall, it’s elevated off its roadbed rails by its elongated sides. This allows it cars to pass underneath the bus, or it to overtake cars on the road, regardless of traffic conditions. Check out the 2012 concept video, still in play, here
Downsides include the fact that only personal-sized vehicles, like cars and SUVs, could fit under the bus — trucks will have to find another route. Also, as BoingBoing’s Cory Doctorow points out , the concept video fudges the bus’s physics: where a real-live straddling bus’s turn radius would make it impossible to corner at most intersections, the artistic rendering conveniently bends parts of the bus that shouldn’t, to make it work. Both issues are major (ahem) roadblocks to real-world use.
But Song Youzhou is already addressing some of the problems: the new physical model employs more articulations to make those pesky turns easier. However, only time (and more prototypes) will tell if the “land airbus” will ever take to the streets. Until then, we can enjoy the daydream of a peaceful, traffic free glide in to work — if our work isn’t already in our living rooms, that is!