Thoughts from the SXSW Partner Sessions

I was lucky enough to attend some of the Partner Programming Sessions during SXSW on Monday 3/14/2016 and Tuesday 3/15/2016 and these are a collection of thoughts/topics/points that I found interesting.

Monday – 3/14/2016

Swipe Right: Get Intiamte with Implantable Devices, Biometric Tattoos and Biosurvellance

  • There is mountains of data out there. Drilling companies will sample drill bit rates at 60 times/sec. Not all data has value but it can be combined with other data sets for trends and information about populations rather than individuals and that has enormous value.
  • 4 A’s of Data: Access, Aggregate, Analyze, Act. – There is a ton of data out there now from sensors (Access), if we combine the data with other data sets (geography, environment, population, etc) we can see much more (Aggregate), then we need to look at the data and do the analysis (Analyze), and finally we can act accordingly and make smart decisions.

Leveraging an Open Source Ecosystem for App Development

  • The community is the starting point for open sourcing, start with open communities on your road to an open ecosystem. – Don’t recreate the wheel or start in a closed system.
  • Open Data Repositories (api.data.gov, Government makes it available via APIUmbrella, the NASA datasets and APIs are crazy awesome https://data.nasa.gov/developer).
  • For Netflix OSS allows them to deploy to 150 countries in one day – That’s amazing!

Building Secure Products in a Connected Society

  • Finding a vulnerability is an art; diversity is key to the hunt. – There is no direct procedure so you need people of different skill sets, backgrounds, cultures to approach the problem in different ways.
  • People like availability. – Such a simple statement, and so true, if Netflix were to suddenly turn off streaming people would lose their minds! Once people have a service and get used to it being available, it’s continually availability is expected and almost demanded.
  • Not all data is created equal. Some is finite and some has value at certain times. There is value inside and outside the organization, classify data based on that.
  • Raise adversarial cost higher than the value of the target. Adversaries are not going to spend $1 million to get a $5 target.
  • Abstract away the data if you can! – Makes things easier to start this way or have it be the norm, especially around HIPAA, PII.
  • Build in security from the ground up on all the layers. On the hardware layer check values are in the correct range, decide that if not it will fail open, or fail closed, or some combination. Regardless notify someone!

Tuesday – 3/15/2016

Harness Data Science Superpower for Global Good.

  • Anyone can start as a data scientist, it doesn’t require specific schooling. With mastery comes mathematical models and advanced understanding, but finding trends in an Excel workbook is data science.
  • If the data contains numbers than trends and patterns are key. If the data is more abstract like images or video then creativity and ingenuity are key.

 

Like all things SXSW the sessions I got to see were great, they got me thinking on all sorts of stuff and I wish I was able to attend more.

Leave a Reply

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