We introduce Running with Scissors, who we are, and why we do the series - to talk about AI in an easy to understand way, to bring AI into the real world in ways other companies aren't, and to show how useAIble stacks up to other ML engines in different scenarios.
Mad nerds out about Logistics and Supply Chain management...Guys then tease me a bit, but realize it's a meaty, real world application. Mad introduces the Beer Game (background, beer jokes, etc) and overview of our Logistics simulation. Cut in to show visual of Beer Game from open source website, discuss how complex inventory management is throughout the supply chain and eliminating out of stocks is important. A hard problem for standard AI....
Lay down the challenge
Here we take turns explaining how hard this is, and how much of a challenge it will be. We describe what engine we're stacked up against. We also show the UI now and outline the rules. Overlay the rules as bullet points on the screen, or cutaway and we talk about the on-by-one revealed bullet points.
Show TensorFlow first
Victory Lap and Wrap Up
Ok, so now we're narrating a GTF (google tensor flow) session, or an edited series of sessions. Watch it fail, maybe improve a little bit, etc. End of each session has final score in big print. We create a leaderboard of the top 25 scores, and a total of the sessions GTF has completed, including total elapsed time. Maybe we should play it too...dunno.
Now we show uAI! Same gig, show it failing, getting better. Showing total elapsed time. Show gradually how the leaderboard gets full really quickly of uAI. We need to show how we solve it better, shorter time, higher scores, better optimization.
We emphasize that AI/ML is hard in the real world once you get out of standard predictable scenarios...Emphasize how practical this is. Click the links if you want the open source code on GitHub - transparency, etc....Maybe we pop open a cold one...Dew for Rix, canned beers Brandon and I since we won the game. See ya next time!