Churchill Club Top 10 Tech Tech Trends Debate 2013 — I chose Deep Learning / Machine Learning

We debated the top 10 tech trends last night. Each of us came up with two that were meant to be non-obvious and potentially explosively important over the next five years (this is really hard, and a bit stressful, as I have learned trying to come up with two ones each year).

After each of us defends our trend, the VC panelist hold up a paddle, either green for agreement or red for disagreement, and hash it out. Then the audience votes by phone. Trends and results and more photos below.

It was a great group this year with fun repartee. You can read the recap by Forbes or watch a video of the whole event (I revisit my 2012 predictions at minute 4:43, present trend 3 on machine learning at 29:15 and trend 8 on the rich-poor gap at 1:20:14).

Here are the trends, with the audience vote in brackets. The panelists split two red and two green on every trend except the last one.

1. EyePhones Replace iPhones [David Cowan, 60% disagree]
“Terminator Vision” enables new classes of applications that drive mobile computers closer to our cognitive pathways. Eyewear computers log our lives for robust searching, sharing and diagnostics.

2. The “Right Now” Economy [Venky Ganesan, 54% disagree]
Proliferation of an ubiquitous user device (“mobile”), presence of consumer web services (“cloud”), and emergence of significant analytic abilities (“big data”) moves today’s consumers from a “plan ahead” economy to a “Right Now” economy.

3. Deus Ex Machina: Machine Learning Innervates the Tech Frontier [Steve Jurvetson, 75% agree]
Machine learning is the technology under the covers that powers many of the exciting new products that leverage big data to appear nearly magical. Imagine a Google Research approach to everything.

4 .The Individual Revolution [Alfred Lin, 66% disagree on obviousness]
Technology has given us indispensable tools to unlock our personal and professional potential, and enabled all of us to become global entrepreneurs. Welcome to the individual revolution.

5. US is the Supreme Cyber Security Force in the World and its Primary Force [George Zachary, 63% agree]
Horrific, unchecked physical terrorist incidents drive fear; citizens accept complete observation by the functions of a police state. A devastating electronic attack results in govt. militarization of major gateways and backbones of the Internet.

6. Cyber Warfare Becomes a Good Thing [David Cowan, 56% agree]
Cyber missions are instant, effective, relatively free, and non-lethal. Advantaged technologically, America starts dismantling conventional military forces, replacing them with cyber capabilities, completely disrupting the Defense Industry.

7. Certifications, Not Diplomas [Venky Ganesan, 50/50 split]
The emergence of MOOCs and other lifelong learning educational sites e.g., Coursera, Udacity, etc. on the Internet means that the future is defined less by where you went to school but more by what you know. The educational world becomes flat.

8. Erasing the Digital Divide Ironically Accelerates the Rich-Poor Gap [Steve Jurvetson, 70% agree]
Technology democratizes upward mobility and raises the bottom of the pyramid but stretches it into a conical spike—where an ever-shrinking percentage of people control the info economy embedded with winner-take-all network effects and power laws.

9. Personalized Medicine [Alfred Lin, 56% agree]
The ability to cheaply sequence genomes and analyze big data affects every aspect of medicine as we know it, from the treatment of cancer, to drug selection, to the engineering of food.

10. Wearable Computing is the Watch, Not the Glass [George Zachary, 59% disagree as well as the four panelists.]
The watch supplants the phone as primary display & interaction interface of the phone. Initially, it is focused on communications, messaging, social networks. The Swatch of the future is a S/W programmable watch. Health functions added over time.

(top photo by Ed Jay)

11 responses to “Churchill Club Top 10 Tech Tech Trends Debate 2013 — I chose Deep Learning / Machine Learning”

jurvetson

May 25, 2013 at 3:51 am

VIDEO of my machine learning prediction.

Defending my predictions

Goofing on Google Glass not being distracting while Alfred from Sequoia is speaking…

The votes

(it looks like they cut off trend #4 off the bottom of the vote as it was a bit confused)

And the award for top two… a dunce’s cap or wizard hat?

And sure enough, someone in the audience had a 3D-printed pair of immersive glasses with electrochromatic glass and embedded camera…

and an quadcopter they can control with outstretched hand gestures… looking like a Jedi…

Reply
Chucil

May 27, 2013 at 3:31 am

Sorcerer’s hat —

"To the lonely
Corner, broom!
Hear your doom.
As a spirit
When he wills, your master only
Calls you, then ’tis time to hear it."

Reply
jurvetson

May 28, 2013 at 6:45 pm

Chuchil – this is what makes Marvin the robot horribly depressed… =)

Meanwhile, topical to scary trends 5 and 6, from the Washington Post exposé today (with links to my photos):

“Chinese hackers have compromised the designs of some of America’s most sensitive and advanced weapons systems—including vital parts of the nation’s missile defenses, fighter aircraft and warships… Also compromised were designs for the F/A 18 fighter jet, V-22 Osprey, F-35 Joint Strike Fighter, UH-60 Black Hawk helicopters and the Navy’s new Littoral Combat Ship meant to prowl the coasts.”

When I think of combative futures in a milieu of accelerating tech change, a key issue is the pace of progress in defensive and offensive modalities. We sometimes imagine super-powerful offensive capabilities of the future versus present day defensive capabilities. Clearly both will advance, but there may be extremely vulnerable windows temporally along the way, and perhaps an argument can me made for certain technologies being inherently offensive where the potential for chaotic damage and entropy vastly outstrips the ability of the same technology to maintain order (e.g., bioweapons). And some technologies are enabling ever-smaller groups, soon individuals, to design and deploy WMDs (certain cyber attacks, genetically modified pathogens) and some simple attacks need no technology or skill (e.g., someone road-tripping through California in the heat of summer + Santa Ana winds with a box of matches and some gasoline -> overload firefighting resources systemically).

Offensive cyber-code and autonomous agents today are not so different from the bio and then nano-weapons of tomorrow. The cell is but a vessel for the transmission of code. So I think humanity will cut its teeth on cultural norms and responses (police state, cyber-counter-guerillas (beyond governments to posses and bounty hunters), and a societal immune system for the crazy ones) in response to the imminent cyber threats… and then we will face bio threats… and finally nano threats. So there is little reason to focus on the latter until we have solved the former.

When I discussed some of this with Newt Gingrich, of all people, in 2004, it prompted him to do a bit of research on his own, and he concluded: "Biological warfare, biothreat, is the largest threat to the human race, a substantially bigger threat than nuclear war."

P.S. Some details on genetically-modified pathogens, my second blog post ever (here and here):

At perhaps no time in recorded history has humanity been as vulnerable to viruses and biological pathogens as we are today. We are entering the golden age of natural viruses, and genetically modified and engineered pathogens dramatically compound the near term threat. Bill Joy summarizes that “The risk of our extinction as we pass through this time of danger has been estimated to be anywhere from 30% to 50%.”

Why are we so vulnerable now? The delicate "virus-host balance" observed in nature (whereby viruses tend not to be overly lethal to their hosts) is a byproduct of biological co-evolution on a geographically segregated planet. And now, both of those limitations have changed. Organisms can be re-engineered in ways that biological evolution would not have explored, or allowed to spread widely, and modern transportation undermines natural quarantine formation.

One example: According to Preston in The Demon in the Freezer, a single person in a typical university bio-lab can splice the IL-4 gene from the host into the corresponding pox virus. The techniques and effects are public information. The gene is available mail order. The IL-4 splice into mousepox made the virus 100% lethal to its host, and 60% lethal to mice who had been vaccinated (more than 2 weeks prior). Even with a vaccine, the IL-4 mousepox is twice as lethal as natural smallpox (which killed ~30% of unvaccinated people). The last wave of “natural” human smallpox killed over one billion people. Even if we vaccinated everyone, the next wave could be twice as lethal. And, of course, we won’t have time to vaccinate everyone nor can we contain outbreaks with vaccinations.

Imagine the human dynamic and policy implications if we have a purposeful IL-4 outbreak before we are better prepared…. Here is a series of implications that I fear:
1) Ring vaccinations and mass vaccinations would not work, so
2) Health care workers cannot come near these people, so
3) Victims could not be relocated (with current people and infrastructure) without spreading the virus to the people involved.
4) Quarantine would be essential, but it would be in-situ. Wherever there is an outbreak, there would need to be a hair-trigger quarantine.
5) Unlike prior quarantines, where people could hope for the best, and most would survive, this is very different: everyone in the quarantine area dies.
6) Where do you draw the boundary? Neighborhood? The entire city? With 100% lethality, the risk-reward ratio on conservatism shifts.
7) How do you enforce the quarantine? Everyone who thinks they are not yet infected will try to escape with all of the fear and cunning of someone facing certain death if they stay. It would require an armed military response with immediate deployment capabilities.
8) The ratio of those available to enforce quarantine to those contained makes this seem completely infeasible. With unplanned quarantine locations, there is no physical infrastructure to assist in the containment.
9) Once word about a lost city spreads, how long would it take for ad-hoc or planned “accelerated quarantine” to emerge?
10) Once rumor of the quarantine policy spreads, doctors would have a strong perverse incentive to not report cases until they made it out of town…

Reply
Chucil

May 29, 2013 at 7:39 am

=)

Reply
jurvetson

Jun 11, 2013 at 6:23 pm

I just read the BusinessWeek article on the U.S. offensive:

”Created in 1952 to intercept radio and other electronic transmissions—known as signals intelligence—the NSA now focuses much of its espionage resources on stealing what spies euphemistically call “electronic data at rest.” These are the secrets that lay inside the computer networks and hard drives of terrorists, rogue nations, and even nominally friendly governments. When President Obama receives his daily intelligence briefing, most of the information comes from government cyberspies, says Mike McConnell, director of national intelligence under President George W. Bush. “It’s at least 75 percent, and going up,” he says.

“You’re not waiting for someone to decide to turn information into electrons and photons and send it,” says Hayden. “You’re commuting to where the information is stored and extracting the information from the adversaries’ network. We are the best at doing it. Period.”

The men and women who hack for the NSA belong to a secretive unit known as Tailored Access Operations. For years, the NSA wouldn’t acknowledge TAO’s existence. A Pentagon official who also asked not to be named confirmed that TAO conducts cyber espionage, or what the Department of Defense calls “computer network exploitation,” but emphasized that it doesn’t target technology, trade, or financial secrets.

According to one of the former officials, the amount of data the unit harvests from overseas computer networks, or as it travels across the Internet, has grown to an astonishing 2 petabytes an hour —that’s nearly 2.1 million gigabytes, the equivalent of hundreds of millions of pages of text.”

Cyber-defense may be very different than cyber-offense. Some argue that open disclosure of defense modalities can make them stronger, but offensive tactics need to be kept private for them to be effective. This leads to a lack of transparency, even in the chain of command. This leaves open the possibility of rogue actors, or simply bad local judgment, empowered with an ability to hide their activities. We may suspect that this is already happening in China, but why would we expect it not to happen elsewhere as well?

Reply
jurvetson

Jun 13, 2013 at 4:32 am

and much discussion ensued

Reply
301man

Jul 10, 2013 at 5:32 am

Steve Jurvetson: Thank you for the photo of you wearing Epiphany Eyewear. I’m including it on the Epiphany Eyewear Article page on Wikipedia and will attribute the photo to you. Thank you!

Reply
301man

Jul 10, 2013 at 5:43 am

Steve Jurvetson:

Here’s the link to the article: en.wikipedia.org/wiki/Epiphany_Eyewear

Here’s the link to the photo: en.wikipedia.org/wiki/File:Steve_Jurvetson_Wearing_Epipha…

Thank you!!!!

Reply
301man

Jul 14, 2013 at 2:40 am

Mr. Jurvetson: I located the author of the Epiphany Eyewear image and he has approved its use on Wikipedia. Sorry if there’s been any confusion. I originally thought you were the photographer. Thank you for your understanding.

Reply
jurvetson

Nov 8, 2015 at 7:20 pm

Where it all started for me…. Rumelhart’s neural network class in 1987

Reply
jurvetson

Mar 28, 2019 at 11:18 pm

And this just in: Congrats to Yoshua Bengio, Yann LeCun and Geoffrey Hinton on winning the ACM Turing Award — for their work with neural networks!

“What we have seen is nothing short of a paradigm shift in the science. History turned their way, and I am in awe.” — Oren Etzioni, the chief executive officer of the Allen Institute for Artificial Intelligence in the NYT today.

Reply