This is the second of three posts covering the fourth annual GeekWire Summit, held October 2 – 3, 2015. GeekWire Summit 2015, like its predecessors, featured talks and panel discussions from leading figures in technology and business in the Pacific Northwest. You can read the first part of our coverage here.
This year, the conference had a specific focus on the future of business and technology, particularly looking at current technology and its possible applications in the future. Here are some of the highlights from this year’s GeekWire Summit.
The crisp autumn morning brought technology enthusiasts from all over to listen to one-on-one interviews with local entrepreneurs and technology leaders in the Pacific Northwest, and the format provided comfortable access to notable speakers’ opinions and observations. Panelists were asked to give their personal forecasts on what technologies would be relevant in 2018, and some distinct trends appeared, particularly self-driving cars, the power and potential for mobile devices, and the emergence of augmented, virtual, and blended realities into our daily lives.
On Amazon’s Limitless Ambition
A panel on the first day of the conference featured former Amazon leaders weighing in on the recent high-profile piece published in the New York Times, where employees were interviewed about Amazon’s internal expectations of employees and the corporation’s intense culture.
In a panel moderated by a contributor to the original Times’ piece David Streitfeld (@DavidStreitfeld), former Amazon employees Sandi Lin (@sandislin), Dave Cotter (@davecotter), and Nadia Shouraboura (@NShouraboura) commented on the the article’s portrayal of Amazon’s practices and culture and their own experiences.
For the most part, the panelists avoided being critical of the culture and structure that generated so much interest in the article, and it was clear that the panelists felt they owed much to Amazon. All three panelists are currently involved in their own startups, and said they took inspiration from Amazon’s core values to help inform their current ventures. They felt that working at their own startup companies was much more difficult than working at Amazon, whose structure provides a safety net even for its more experimental ventures.
Though the discussion largely skirted around more sensitive topics and critical questions, it was hard to miss the sense that the speakers felt they were on the defensive. Cotter, in particular, took issue with the use of the word “intense” to describe Amazon’s culture. There was an extended collective effort to clarify that “intensity” was not a negative term, but simply a description for the unique pressures and demands within Amazon’s ever-innovating environment, which Streitfeld compared to the atmosphere of a startup.
Another tense moment was during the Q&A portion of the panel when Lin described how Amazon has a culture which naturally attracts top talent, and that such talent naturally works hard. Lin’s statement that “bosses and managers aren’t responsible for burnout” of their employees elicited an impassioned “Yes you are!” from an audience member. Though Lin did not respond directly to the audience member, the discussion afterwards focused on how the panelists handle the responsibility of leadership now that they are at the head of their own companies.
At the conclusion of the panel, Streitfeld shared the reason for his own fascination with Amazon and why it remains an interesting company to watch: “It always escalates what it’s doing. Its ambitions are never satisfied. Amazon’s ambitions are limitless, and no company has survived with limitless ambition.” And yet despite that, Amazon has continued to thrive, perhaps proving that it has tapped into something not easily replicated.
Changing Expectations Drive Civic Ventures
GeekWire co-founders John Cook and Todd Bishop engaged venture capitalist Nick Hanauer (@NickHanauer) on stage during the first day of the conference in a discussion that covered issues concerning global health, social media and social/civic balance as well as the potential future of automated vehicles.
Hanauer urged the audience to “be civically engaged,” and not only in times of crisis. He pointed out the challenge facing Amazon, as a large company, is to be more civically involved in the community. He noted that the NYT article on Amazon might only be news because “we’ve changed our expectations” of large companies, and we now expect them to be supportive and mindful of the effect that employment has on a workforce.
However, he was quick to add that the onus to do more for the community is not something that lies on large companies alone. Rather, anyone doing something that could potentially disrupt existing structures needs to be more civically minded. “There’s a significant problem in the way new business models inevitably disrupt the economy. We have to find a [better] way so that they are not civically, economically, and socially disruptive,” Hanauer said, later adding, “We have a responsibility of leading the charge to find ways to mitigate the disruptions [the technology community] creates.”
Hanauer also spoke passionately about the importance of the $15 minimum wage and why tech companies should care.
A final reflection from Hanauer’s chat was that along with biotech taking the drivers seat in the contemporary technology of 2018, that the advances of artificial intelligence and driving cars will propel our society forward. Hanauer notes that his 15 year-old son may be of the last generation of humans who need to learn how to drive a car themselves. Overall, Hanauer made clear that the idea of automated cars taking the wheel from teen drivers is an idea that’s worth driving around the block a few times.
Self-Aware Cars and the Future of Artificial Intelligence
In a third panel on the first day of the conference, John Markoff (@markoff), science reporter for the NYT, was joined on stage by Peter Lee (@peteratmsr), global head of Microsoft research. They also discussed self-driving cars, but more broadly artificial intelligence and its contemporary challenges. Coming into the discussion with two different points of view, they attempted to give the audience a more well-rounded picture of what’s next for AI.
Markoff and Lee both agreed that one of the biggest challenges facing self-driving cars right now is the hand off problem. This refers to when an automated driving situation becomes too complex for the AI to safely handle the vehicle, and it “hands off” control to a human inside the vehicle. Lee noted that an obvious drawback was the lack of situational awareness the human driver might have available on such short notice. A driver may be distracted by the view of the scenery, a game, or even be sleep, and therefore unable to disengage in time to take control of the wheel. According to the panelists, a safer option might be simply to instruct the car to pull over and park when it is overwhelmed.
That thought certainly seemed to hold true when GeekWire attendees were shown a video reel of the results of the DARPA Robotic Challenge, which aimed to create robots that were capable of being operated in places inaccessible to humans. Sadly, even with the best funding and the best engineering teams on the task, common and routine actions are still incredibly difficult for robots to mimic, including tasks as basic as opening a door, hold a drill, or “simply” exiting a vehicle.
— Jim Loter (@jimloter) October 1, 2015
Lee also brought up the new perspectives that data scientists have after looking into the wealth of information that can be garnered from “deep learning systems” gathered from immense data sets. Broadly, deep learning refers to an approach to AI learning that uses a hierarchy of concepts to allow computers to build complex ideas out of simple ones. The panel explored the potential of deep learning and how it may contribute to the development of future systems, although that benefit may not be realized anytime soon. Attendees were encouraged to reflect upon Paul Saffo’s (@psaffo) line that one must “never mistake a clear view for a short distance,” and there is much learning to do in the AI sphere.
One type of discovery in this type of data is particularly useful in understanding the user at an emotional level. Lee noted his personal experience, saying “people ask for help when they’re stressed out.” The implication is that through an automated contextual analysis of data in a particular time frame, we can learn more about the factors that contribute to the issue. Simply put: the future of data analysis is the understanding the head space of a user.
Another challenge faced by AI is the rudimentary understanding of seemingly simple symbols in context. For example, an AI system might be able to successfully identify a simple clipart arrow as an arrow, but the relevance of the arrow in context to the surrounding content is still difficult for it to ascertain. That takes careful, complex programming.
When asked about what technology they perceived would be most relevant in 2018, Markoff noted that major opportunities lie in deep learning, an approach to AI learning that uses a hierarchy of concepts to allow computers to build complex ideas out of simple ones. Lee predictws serious advances in deep learning, particularly with human discourse, as well as an end to Moore’s Law.
Additional coverage and analysis for this article was provided by Samantha Hautea. Stay tuned this week for our continued coverage of Geekwire Summit 2015. You can find our previous coverage here.