Afraid to Fail

Over the weekend I was referred to the NY Times article on Google X, a secret lab where where Sergey Brin and other Google employees tackle important projects that aren’t yet ready for primetime. And when they say important they mean universally, as in universe, important.

Take for instance the elevator to space project, or the one that the laser scanning industry is more directly connected with the driverless car. The founders of Google have it as their mission to solve really BIG problems using technology. I think these 2 will qualify, although I am not quite sure about the elevator concept.

The key insight I think comes from Google co-founder Larry Page, “I just feel like people aren’t working enough on impactful things. People are really afraid of failure on things, and so it’s hard for them to do ambitious stuff.” I think this is the case in most companies that I know. This kind of culture just does not exist, unless you have the resources of a Google.

So where will the next breakthrough innovation in laser scanning be developed? In a university lab? Maybe, but their track record is not good for doing this kind of work. How are we going to create that environment where it is O.K. to take risks and to fail? Any ideas?

This entry was posted in Mobile LiDAR, Technology. Bookmark the permalink.

3 Responses to Afraid to Fail

  1. Casey McLaughlin says:

    Good question and one I struggled with in school. I always thought there should be a journal of failures. Learning from mistakes is something we should be good at but I’m not sure the current market drivers or the grants system is supportive of the investigative nature of Google X. Are there other examples out and about? What about tales of success?

  2. Daniel Huber says:

    The academic research environment is actually the ideal location for investigating risky and potentially high impact ideas, and academia has a long track record of successes in innovation. The academic environment is more tolerant of failure than industry because academic research does not have to make a profit, and academic researchers have the flexibility to adapt rapidly if roadblocks are encountered. Furthermore, the open and free nature of academic research encourages learning from the mistakes and successes of previous experience.

    It should be noted that all of the leaders of the Google Car project and over half of the engineers are former faculty or students of CMU’s Robotics Institute, and nearly all of them were involved with the CMU or Stanford teams on the DARPA Urban Challange (as well as the DARPA Grand Challenges). The technology that enables the Google Car to be successful is a direct descendant of the methods developed and fine tuned by the university teams in the DARPA challenges. These methods were, in turn, built on previous research on autonomous vehicles that was conducted in a university setting. Our department even had a self-driving car 16 years ago. Take a look at http://www.cs.cmu.edu/afs/cs/usr/tjochem/www/nhaa/nhaa_home_page.html.

    Another example is the Kinect. The key capabilities that enable the Kinect to recognize human pose in real time were developed by academic researchers who have been hired by Microsoft, and those ideas were themselves built upon the experience of previous academic research, including the failures. The stereo algorithms embedded in the Kinect are build on years of research on stereo computation.

    Even the idea of Google itself originated in an academic project, and many of the key ideas that allow Google’s technologies to work originated or are built upon academic research.

    To claim that people are not working on “impactful things” is ridiculous and untrue. While it may be true that industry has reduced its internal R&D funding, that is not too surprising given the tough economic times of today. Still, the academic research environment continues to be productive and innovative. Each year, we see amazing new advances in technology, many of which have significant impact on the world.

    Most of the funding agencies that sponsor academic research have an explicit goal of funding highly innovative projects. For example, DARPA has the concept of “DARPA hard,” which means that a problem has to be nearly unsolvable before they consider it interesting. NSF has a similar criteria requiring any proposal to have significant intellectual merit and broader impacts.

    Many research projects are successful, and many do not achieve their original goals. No one wants to classify their project as a “failure,” and usually, researchers will find some success even when an initial idea does not work out.

    One problem with failures in academic research is that it is much more difficult to publish a paper about what doesn’t work than it is to publish a paper about a success. Nevertheless, many papers include information about approaches that did not work as well as the proposed approach, and much of the knowledge about what doesn’t work is communicated between researchers directly.

    As our government continues scale back funding for research, the question is not how are we going to create an environment where it is ok to take risks and fail, but rather, why are we not doing a better job of supporting the excellent environment that we already have in place?

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.