TheVORTEX

As I mentioned before, I signed up for the 160,000-student online Introduction Artificial Intelligence by Stanford illuminati Sebastian Thrun and Peter Norvig. I audited by watching all the videos and doing the little online quizzes. Others did all the homework, took the tests, and wrote code to go with the lessons. Waaay beyond my ancient math skills. One thing I learned is that AI is all about probabilities and algorithms to calculate them for anything in the world you can measure.

Why I bring this up again is that I see where MIT—which has had a lot of packaged courseware online free to the world for a long time—just launched a new program called MITx. MITx is online courses where video lessons, quizzes, material and grading all done automatically. It’s free, worldwide and anybody can sign up. The first class is Circuits and Electronics and it starts March 5.

Is it a coincidence that MIT came out with a competing free online system? I think not. Sanford and MIT are arguably the top two engineering schools in the country. They are in heated competition of have the best “brand” for engineering and innovation. So, bang!, just like that, here’s MIT with outreach worldwide for the leadership title. It harks back to the way Stanford and MIT competed in the 2007 DARA Grand Challenge for autonomous vehicles. Stanford’s AI team under Thrun kicked MIT’s butt.

But there’s another reason to set up these courses. In December last year, not long after the AI course was over, I saw on YouTube a discussion session between Sebastian Thrun and some people at a Singularity University seminar. In the session Thrun said something that hit me between the eyes. He said after the final exam they had ranked all the students that completed the course and found something like 2,000 students that had nearly perfect scores. Over 200 had perfect scores; not one of them was a Stanford engineering student. Most were not in the United States. (Don’t quote me on the exact numbers. I’ve looked high and low on YouTube and can’t find the video I saw to double-check. No luck.)

Given the poor performance of American students in science and math in many international standardized tests this wasn’t exactly earth-shaking. But what got me was that Stanford engineering and his team had a list of many of the best and brightest people in the world in math and programming! That is like being the only pro basketball scout in the country to have ever seen the top scoring college players in the country. What’s a list like that worth? To the Stanford engineering department? To recruiters at Google, Apple and other Silicon Valley tech firms? The AI course turned into a fantastic screening tool to find and potentially recruit the next generation of top engineers, scientists, and programmers. I can’t be the only person to think of that. The folks at MIT must have slapped their foreheads hearing that like I did.

The latest shoe to drop: Sebastian Thrun is leaving this job as tenured director of the Stanford AI program to join a startup called Udacity where he’ll be teaching an online worldwide course: CS373 Programming a Robotic Car. (Watch out! Next time I take my ‘99 Civic out for a spin I may be sitting in the back seat.)

Thrun as expressed interest in innovating in mass education, but is that the only reason to take a left turn in his career? I mean, he’s the rock-star of AI. I suspect that he and Stanford had some sort of parting of ways. Free mass education may not have been the University’s idea of how to conduct a 100+ year-old university.

I’ve been posting about robotics and AI lately. From robotic cars and tractors, I’ve turned to where the impact of AI may be just as big but less expected: in white-collar and “informaitn worker” jobs. Part 1 was about computing moving into financial market transactions. This time, lawyers.

A NY Times article by John Markoff recently looked at software in the legal industry. Systems and software (think IBM’s “Watson”) are being put to use like a backhoe where troops of lawyers usually do the pick-and-shovel labor of discovering relevant documents for cases.

Another e-discovery company in Silicon Valley, Clearwell, has developed software that analyzes documents to find concepts rather than specific keywords, shortening the time required to locate relevant material in litigation. Last year, Clearwell software was used by the law firm DLA Piper to search through a half-million documents under a court-imposed deadline of one week. Clearwell’s software analyzed and sorted 570,000 documents (each document can be many pages) in two days. The law firm used just one more day to identify 3,070 documents that were relevant to the court-ordered discovery motion. […]

Quantifying the employment impact of these new technologies is difficult. Mike Lynch, the founder of Autonomy, is convinced that “legal is a sector that will likely employ fewer, not more, people in the U.S. in the future.” He estimated that the shift from manual document discovery to e-discovery would lead to a manpower reduction in which one lawyer would suffice for work that once required 500 and that the newest generation of software, which can detect duplicates and find clusters of important documents on a particular topic, could cut the head count by another 50 percent.

Indeed, immediately after “Watson” pulled off the Jeopardy take-down, IBM announced it was going to market Watson-like systems for medicine (to help docs make better diagnoses by plowing the literature) and for law as outlined above. Lawyers don’t get a lot of sympathy, but I have the sense tools to reduce the hours worked by legal functionaries comes from outside the profession, not inside. But it’ll be an interesting test case to see how these expensively-educated and well-paid people adjust to system lightening their load and perhaps their paychecks. File an injunction maybe?