Get email delivery of the Cadence blog featured here
Jim Hogan is a graduate of San Jose State University. Despite the prestige of Stanford and Berkeley producing a lot of researchers, SJSU is a major resource producing the large cadre of practitioners that Silicon Valley requires. Last week, at SJSU, Jim presented his thoughts on what he calls the fourth industrial revolution, what he calls "The Cognitive Era." He repeated the presentation the following evening, as the dinner keynote at the electronic design process symposium, which I wrote about on Monday. See my post on Solving the Design to Manufacturing Problems in Milpitas. (In addition, he will be coming to Cadence on November 15 to give the presentation as part of our Distinguished Speaker series, so if you are a Cadence employee you can attend or watch the live-stream or the recording).
I gave Jim a call to get a bit more color, not so much on the cognitive era, but what he is trying to achieve at San Jose State.
If you look at the title of this post, you might think that SJSU has a school of cognitive science. But in fact, it doesn't...not yet anyway. And that is something that Jim is hoping to change. Universities in general, and SJSU in particular, are not very good at dealing with disciplines that cut across departments. I remember, for example, when I was at Edinburgh that we had separate departments of electrical engineering, computer science, and artificial intelligence. These three departments barely talked to each other. Electrical engineering and computer science didn't work together at all. (In fact, when we started teaching IC design in the computer science department, it seemed to be considered a declaration of war).
Many of the departments at SJSU teach courses on little pieces of the problem: data structures in CS, matrix algebra in math, social science is doing lots of deep learning on big datasets, work on facial expressions, and more. But Jim feels it needs its own school. He draws an analogy to when he went to SJSU forty years ago. There was no school of computer science, you had to study math and then take some courses in engineering, and hand-craft your own computer science degree. Now the school of computer science, at least in terms of the number of graduates, is bigger than engineering. Jim feels that cognitive science will be like that in a decade. If you didn't read yesterday's post on Breakfast Bytes Machine Learning for Higher-Performance Machine Learning, then you might want to do so. It is a survey of all that Google is up to in the Google Brain team, everything from how a songbird's brain is wired up, to building chips for petaflop level inference. Jim expects the school of cognitive science to be similar, doing everything from designing specialized hardware, to working on new paradigms for how humans should interact with computers, to linguistics, and even philosophy.
Jim has been pitching this idea for seven or eight months now, and there is very little pushback. It is a good time for change too. SJSU has a new president, and a new dean of the school of engineering. If you know anything about academics, you know they will want to make their mark during their tenure, and they are supportive of the program. SJSU is also building a new $65M innovation center building, breaking ground early next year. The school of cognitive science will be in that building. One immediate challenge for SJSU is to define exactly what cognitive science is. UC San Diego has a school of cognitive science but that may not be the best match since UCSD is more of a research university with a different mission than a state school like SJSU. Jim says that the next meeting in his series will probably be a panel session in February about what the definition should be for SJSU, presumably including some of the faculty who have put some effort into the problem.
Now, on to what Jim said that evening...
There have been four main eras, kicked off with their own revolutions. First, the agricultural revolution, around 10,000 BC, leading to agriculture-based society. Jim, being a computer scientist, calls this 0.0 (hey, chip designers number their buses from 0, too). This seemed to have happened independently in different parts of the world.
Then, around 1750, starting in England, was the industrial revolution leading to a society based on mechanical things (such as looms) driven by water and steam power. This was revolution 1.0.
Next came 2.0, and electricity, in the late 19th century, leading to electric-powered mass production.
The next revolution was electronic. You can choose your date for this, but computers and integrated circuits (and software) were the bread and butter of revolution 3.0.
Revolution 4.0, the one Jim was talking about, is starting roughly now. Physical systems, such as machines and robotics, will be controlled by automation systems equipped with machine learning algorithms. Minimal input from human operators will be needed.
The above diagram shows expected changes in employment (in thousands) between 2015 and 2020. These are a continuation of a trend that has been going on for a few decades, as automation of one sort or another, has replaced people. In our industry, a modern fab requires very few people to run it, and almost none of them are actually inside the fab itself. But already computers are better than humans at reading X-rays, and doing some sorts of paralegal work. It is not just physical labor that is being replaced, but routine brain-work, too. I've seen predictions that any job that is routine enough that it has large repetitive aspects, will be automated in the coming decade or two. I can't wait for my self-driving car...but then I'm also not a truck driver.
The sectors that look most set to benefit (based on the Economist Intelligence Unit anyway) are healthcare, education, finance, infrastructure, and energy. This doesn't necessarily mean that everyone who works in these sectors has a rosey future. For example, I think it is a disgrace, for example, that education has seen so little innovation. If you walked into a classroom fifty years ago, especially a university lecture hall, it would look pretty much the same as today. Look at how much retail has changed, for example, never mind anything high tech that didn't even exist 50 years ago. There were no big box stores, no Walmart-style logistics, no products from China, no Amazon (obviously), not really even supermarkets. Education, not so different.
There are plenty of startups trying to disrupt some of these segments. I think Jim put Theranos on the list only so he could admit that he was one of the first investors in them. Even Jim doesn't always get out with a nice return on his investment. Talking of investment, having spent his career in EDA, you might think that Jim's best investment would be some far-sighted recognition of some key EDA technology. Nope, it came from BlueLithium, which he got involved with when he met one of the founders in a bar, while waiting for his daughter, and realized they needed help structuring the company since they were clueless. He went on the board and Yahoo! bought them for $300M a year later. Moral: it is possible to pay a lifetime of bar bills by hanging around in bars.
In the agricultural and industrial eras, there were returns to being strong. Being the smartest laborer in the field didn't really help much. Also, pretty much anyone who could work successfully in agriculture (aka almost everyone), could work successfully in manufacturing.
In the electronic era, being strong didn't really help (and, given that many assembly tasks were delicate, it could be a positive disadvantage. Those delicate tasks are/were dominated by women). In the cognitive era, it is hard to see that the returns to being smart are not going to increase even more, and the resulting inequality is going to be a huge challenge for every society. It will also widen the rural/urban divide, with the winners largely in the biggest cities (well, urban areas, Silicon Valley and the Shanghai region are not technically single cities), and the losers in small towns and the countryside.That's one reason that the school of cognitive science encompasses sociology and philosophy, too.
Sign up for Sunday Brunch, the weekly Breakfast Bytes email.