Having a lot of smart people work in parallel is fun but it would be far more productive for all of them to work together.
Here is a crazy thought for these kind of contests, make the first round competitive and award to the winner team. Then make a second collaborative round where everyone who gets a pull request accepted gets invited to see the code in action.
We need the money to go towards helping create open source communities instead of reusing the competitive approach from academia.
"it would be far more productive for all of them to work together" -> not necessarily. It's not very common but there are examples of having multiple teams try to solve the same problem in parallel. It ends up speeding up development - they will come up with different approaches, and the best (or a combination) can be selected.
I just got back from Udacity's student conference. They showed a video of the Udacity self-driving car performing at Level 2 autonomy running code from Udacity students which I thought was very cool. They do have an in house engineering team driving what's probably a majority of the work, but the problems they discussed were discrete problem areas which I imagine scale well.
The company is apparently now a registered self-driving car manufacturer, in addition to being an education company.
29 comments
[ 3.7 ms ] story [ 91.9 ms ] threadI'm located in Vancouver, Canada, have many years of C++/Python experience and been studying ML/DL as a hobby.
Here is a crazy thought for these kind of contests, make the first round competitive and award to the winner team. Then make a second collaborative round where everyone who gets a pull request accepted gets invited to see the code in action.
We need the money to go towards helping create open source communities instead of reusing the competitive approach from academia.
It's somewhat effective.
The company is apparently now a registered self-driving car manufacturer, in addition to being an education company.