A wise analyst once said “data science is 99% counting things, and sometimes dividing”.
Whenever I hear about AI or deep learning in the AF, I can’t help but roll my eyes. Those are at the top of the pyramid! Developing, deploying, and maintaining advanced models must build on a extremely strong and robust lower level data cleaning, aggregation, and pipe-lining. Although, flashy and enticing we aren’t ready for AI in the AF.
IMO large investments must be made in the boring stuff. Database management and integration - cloud storage and process consolidation. Navigating 10+ disconnected and individually contracted database systems is not realistically manageable. Not doing the boring stuff right has a negative effect on our ability to observe, orient, and decide. Furthermore, it doesn’t seem that any of these contractors have any incentive to collaborate. Consolidation means loss of business.
Before, we jump the gun to AI. We need to glimpse() and summarise() what we already have. We need to deepen our knowledge of our current processes - we can’t optimize what we don’t understand. Which brings me back to my first point - counting. The easiest and often greatest insights into a process are usually the result of a simple count.
- How many people are doing xyz?
- How many times is xyz done per month?
- How often does xyz happen?
- etc.
Graduating from USAFA and serving in the Air Force are among my greatest life accomplishments, and I desperately want to contribute to this country which has truly blessed me and my family with opportunity. But I’ll admit, I question if this is best way and every time I read a headline to the tune of “How the Air Force is using AI to …” I feel that much more disconnected from the larger organization.
The gov’t and military have the most motivating challenges for any data scientist. Your work directly contributes to the safety and security of our nation, not to the pockets of shareholders. I believe younger generations feel more obligated to serve in roles that benefit society, and the public sector has the strongest pitch. But we need to be empowered - data is the present and future – can we get a handle on ours?