51 Minutes and 21 Seconds Well Spent
(or at 1.5 speed, 34 minutes)
by: Mark Stadtmueller, VP, Product Strategy, Lucd
It has been known for a while that our brains can process language much faster than people speak (link). That is a great thing about watching lectures on YouTube (where for me 1.5 speed often works). But, at whatever speed, the YouTube video of Harvard Professor James Mickens recent speech is “must see TV” for all of us. Thanks to Dr. Chris Carothers for pointing to it.
My takeaways and thoughts:
People who can combine humor (laugh out loud humor), education, brevity, breadth, and history are so important to educating all of us. (I did watch the video again, not for the content, but just to admire how Mickens did that in this presentation). Mickens description of how Machine Learning works accomplishes all of that.
The discussion comparing Liberal Arts to Computer Science is a worthwhile thought process that seems to be discarded too often these days (i.e. taking the time to perceive how others, who do not think like you, may see what you are doing differently, and they may have a point).
Respect for history. Those of us who work on newer things too often view history as not relevant when most times “there is nothing new under the sun”.
The Egg Drop analogy to what we are doing in Machine Learning is spot on. However, maybe Mickens is too hard on that point. We humans have used trial and error to discover things for a long time.
A lot of Mickens concerns about AI (and IoT for that matter) is their connectivity to the internet and reliance on the internet for data to train on. That is true, but responsibly using AI does not require relying on the internet. In Enterprise AI, businesses have many sources of data without relying on the internet and social data. Bias and Responsibly using that data is certainly still an issue, but Enterprise AI does not have as a requirement, leveraging data from social internet interactions.
Thanks to Professor Mickens for his 51 minutes and 21 seconds. All of it, because his answer at the very end to the question about AI/Machine Learning is that it has a place today and it is needed, but forethought is required.