Day two of IBM Edge 2016 is all done, and the focus has shifted to the individual. Let’s get right to the recap:
One of the more memorable talks during the general session was Hortonworks. They’ve helped a transport company do more than simply track drivers. They assemble and analyze lots of information about each driver, the truck, the current road conditions, and other factors. From there, they apply a risk rating to that particular truck and provide updates to the driver about potential hazards. It reduced their insurance costs by 10%.
Florida Blue shared some insights from their POWER deployments and how they were able to get customers serviced faster. One of the more memorable quotes was:
The best way to get a customer happy is to get them off the phone.
They were able to rework how the backend systems retrieved data for their customer service personnel and cut average phone call durations from 9 minutes to 6.
Jason Pontin came on stage with three technology innovators under 35. They shared some of their latest work with the audience and it was amazing to see the problems they’re trying to solve. Lisa DeLuca introduced her new children’s book that helps to explain technology in new ways:
My first breakout session was Getting Started with Linux Performance on IBM POWER8 from Steve Nasypany. This was a highly informative session and you’ll definitely want to grab the slides from this talk whether you use POWER or not.
Steve dove into how to measure and adjust performance on POWER systems. He also gave some insight on how AIX and Linux differ when it comes to performance measurements. There are quite a few differences in how AIX and Linux refer to processors and how they measure memory usage. He took quite a bit of time to explain not only the what, but the why. It was a great session.
My second breakout was Bringing the Deep Learning Revolution into the Enterprise from Michael Gschwind. He kicked off with the basics of machine learning and how it matches up with the functions of a human brain. He provided some examples of objects that the human brain can quickly identify but a computer cannot.
The math is deep. Really deep. One of the interesting topics was stochastic gradient descent (warning: highly nerdy territory). It measures how well the computer has been trained on a particular machine learning task. The goal is to reduce errors and do less brute-force training with the computer so it can begin working independently. It’s oddly similar to raising children.
My breakouts were cut a little short because I was invited to be on theCUBE! It was completely nerve-wracking, but I had a great time. The hosts were fun to work with and the conversation seemed to flow quite well.
We talked about OpenStack, OpenPOWER, and Rackspace. You can watch my interview below if you can put up with my Texas accent:
We headed outside in the evening for a poolside reception. The weather was in the 80’s and it felt great outside!
Everyone made their way inside to see Train perform live!
The concert was great. They played plenty of their older hits and shared a new single that hasn’t been released yet. We even heard some covers of Led Zeppelin and Rolling Stones songs! Some attendees were dragged up on stage to help with the singing and they loved it.