The Recording
The Panelists
- Kevin Feasel
- Tracy Boggiano
- Mala Mahadevan
Notes: Questions and Topics
CAST() and CONVERT()
Our first topic comes from Tracy, who noticed an odd situation in which using CONVERT() on a dataset ended up about 200ms slower than using CAST(). I didn’t have any good answers, though chat and I speculated on possibilities.
The Future (and Past) of No-Code Programming Languages
After that, we meandered a bit on no-code programming languages, using this article as a jumping-off point. There’s some value in no-code languages but finding a general-purpose, no-code programming language which doesn’t want to make you tear your hair out in frustration as soon as you leave the happy path is like searching for El Dorado.
If you’re on that happy path, then they can be a good way of getting started. Examples I used include SSIS, Informatica, and the Azure ML designer. These are custom-purpose canvases for specific operations. Trying to turn diagrams into code is something that we’ve seen since Rational Rose (and maybe even earlier). It just doesn’t work.
Using AI for Movie Funding Decisions
Our final topic came from this Hollywood Reporter article about Warner Brothers using an ML model to help determine whether to fund movie production. We talked a bit about it, including some parts without snark. Just a bit.