It isn’t every day that I publicly disagree with a man who won the Turing Award and was on a first-name basis with Albert Einstein.
But today I respectfully will.
On January 24, an incredible man named Marvin Minsky died. He was one of the fathers of computer science in general and artificial intelligence in particular. I never met Mr. Minsky, but I share his passion for AI. The concept of superintelligence fascinates me. Yes, we’ve got a lot to figure out. There are so many HAL-like dangers we need to dodge. The work is really fascinating and if you want to read some really interesting work by Minsky take a look at his Society of Mind.
In eulogizing Minsky, the BBC linked to one of the last interviews this pioneer granted—a piece in the MIT Technology Review. Be sure to watch the video; I guarantee it will be four minutes and five seconds you will not regret.
The nit I’d like to pick with Mr. Minsky comes at the halfway point.
After reminiscing about the field’s exciting formative years in the 50s and 60s, when there was “a major new discovery every two to three days,” he goes on to note that growth has not kept that heady pace. He blames big business. “We have to get rid of the big companies,” he says, “and go back to giving support to individuals.”
I respectfully—very, very respectfully—disagree on the first half of his sentence.
As an inventor and engineer myself, I get it. Big business can get in the way of good science. But it doesn’t have to be that way. I joined Cisco because I felt that solving some of these very hard problems would be easiest from inside a company with both tremendous resources and a passion for innovation. Resources and passion for innovation are the key words. Three years later, my experience here proves that hypothesis right. We are using our resources and our commitment to solve some hard problems. Others are too; I am impressed by the real progress in Machine Intelligence made at the likes of Google and Facebook, Apple and even IBM.
Big companies clearly have challenges—as Marc Andreesen has pointed out, capital markets are a big issue in large company innovation. Overrotating to shareholder demands can lead to a focus on optimizing performance at the cost of innovation.. We need to fix that. But I do believe that big companies who have innovation in their DNA are part of the solution—not the problem.
While Minsky talks about brilliant individuals and big companies, he doesn’t mention start-ups and I think that’s a worthwhile omission to mention. I am equally passionate about the role of start-ups. I’ve met with the founders of many AI start-ups, and these are brilliant individuals who want to see their research applied. They are creativity with a job to do. They get it done and they are part of the solution, too.
Speaking in 1967, Minsky famously said: “Within a generation … the problem of creating artificial intelligence will substantially be solved.” Clearly, that didn’t pan out. We need to get back to the innovation cadence that Minsky reveled in—a major new discovery every two to three days.
A combination of brilliant minds, start-ups, and innovation-focused enterprises will be the key.
Cisco’s collaboration and IoT groups are really excited to play their part. We are passionate about the role machine intelligence will play in the workplace of the future. We have big ideas and big plans, and I am going to share those ideas and plans here in this blog over time. Machine intelligence will truly revolutionize the workplace over the next 24 to 48 months. It’s the next big thing to hit the enterprise and Cisco intends to play a leading role in this revolution.
I wish Marvin Minsky had lived another two years to see what we have in mind. I think it would make him proud. I hope I am right.
How do you feel about “thinking” machines in the workplace? Let me know @rowantrollope
Rowan,
There are those in the community that argue that “thinking” machines in the workplace, are not bad in and of themselves, but that we’re not ready for them. Our social, economic and cultural structures have so far to go before we can embrace the technology without negatively impacting 25%-50% of the workforce in a way not seen since the great depression.
Would love your thoughts on this..
https://www.youtube.com/watch?v=7Pq-S557XQU
Do we really need them. Let us think people. Who are we and what do we think we want? Think carefully as what we are told we want, nay need, isn’t necessarily what want, nay need.
I completely agree that Dr. Minsky overstated the need to get rid of big business in science. Business is a catalyst for inquiry and makes science meaningful to the masses. However there is a component of Marvin Minsky’s big business lament which deserves greater attention.
With some exceptions, big business’ involvement in science is typically applied research. Even in those corporate research facilities where some pure research is conducted, there is pressure to tie the research to a future product opportunity. This typically means a timeline for application within the next decade or two. If the research happens to result in deeper, more enduring insights then that is merely icing on the cake. However, “enduring scientific insight” rarely succeeds as the primary research objective when corporate funding is sought. As the applied corporate research dollars and influence have increased, support for pure research, and those who would make it their life’s work, has shrunk dramatically.
Pure research fundamentally differs from applied research in its motivation. Pure research gives little, if any, thought to productization when choosing a line of inquiry. Its search for breakthroughs follows where the pure science points rather than where the opportunity for productization points. Imagine if Issac Newton could only get support to investigate those areas most likely to yield products viable in the markets of the late 17th century. A few more interesting product concepts would’ve been produced but how much longer would it have taken humanity to get to the moon or the microchip?
This is the dilema many, especially young, talented researchers face today. In order to apply their skills they require funding. Most of that funding today comes with corporate productization biases and only secondary attention given to pure science objectives. We live in an age where great products are plentiful and happen every few days. Can we say the same about the health of pure research efforts and the potential breakthroughs which will underlie future generations’ opportunities for better lives? Are we pushing the next Issac Newton to build a better Alexa rather than creating the equivalent of Newtonian Physics for the human mind?
These are important questions which currently, even in an election year, are not getting the attention they warrant. We have some of our best and brighest (e.g. Hawking, Musk, N.D. Tyson…) raising issues about the impacts of mind-science and technology. The need has never been greater for pure scientific investigation into the science of mind (both natural and artificial).
Should we get rid of big business from science, or even reduce its contribution? Absolutely not.
Should we greatly increase support and options for pure research with no productization strings? Absolutely yes!
Thanks Rowan for highlighting this important topic and the work your group is doing in advancing machine intelligence, collaboration support and IoT. A healthy alliance between corporate R&D and pure science is at the heart of humanity’s future. The deeper truths revealed by pure research insure that social and economic benefits of applied R&D can be harvested for perpetuity. It is the recent neglect of pure research’s role in the scientific ecosystem that I believe Marvin Minsky was trying to highlight and thereby instigate this conversation. I hope more corporate and public leaders start getting involved in this topic.
Even now Dr. Minsky continues to further the quest for greater insight and better solutions.
Rest In Peace Dr. Minsky.
Thanks again Rowan for considering the key issues of our time and how they relate to the technologies your teams are creating. Very refreshing.
“Are we pushing the next Issac Newton to build a better Alexa rather than creating the equivalent of Newtonian Physics for the human mind? These are important questions which currently, even in an election year, are not getting the attention they warrant. ”
That is a very nice summary of one of the main points I think Marvin would have liked to get across, thanks Tim
And thank you Rowan for a very thought provoking article, I believe Marvin would have enjoyed very much a discussion with you about these issues.
GREAT
This is a good debate really like the comments.
Thanks Janice. I agree. It’s always great when a post starts a conversation.
I believe that, between big business, start-ups and individuals, there are innovations every 2-3 days; some just haven’t come to light. But, the collaboration among the three of bringing these innovations to light must include consideration for security and privacy. There needs to be significant efforts to self-police to ensure it’s an Internet of Ethics. Cisco was an early adopter of both IoT and IoE – and quickly recognized that IoE/IoT requires also the Security of Everything.
Rowan,
“thinking” machines in the workplace: it is the future. When I’m thinking about what is my job? What is my value added? I can only answer that my job is all about to share intelligence with team. I see easily the introduction of IA in the workplace to replace progressively the part of my job with low value or highly time consuming.
It’s a great idea to bundle IA with Collaborative solutions.
Good luck
I don’t think ai is a good idea in work force because it will create dumber and lazier people where machines think and gives decisions.Also we we don’t have the ingredients on earth to succeed this theme.Its only look good on scify movies…I believe machines that take commands and execute this commands would be ticket.This way you can hold responsible those who gives this commands…..