Read this yesterday and absolutely loved it. Especially feel the pain regarding working with management. I think there's an accountability that comes with evaluating management decisions with data that nobody really wants. I still have a lot of half-formed thoughts/opinions about this but it really feels like data-driven requires strict discipline but the data people who would be accountable for that discipline both 1. aren't empowered to wield it and 2. probably don't want to wield it anyways.
Also agree about the simple tools but it's really hard from a career perspective. If I deploy XGBoost in production and put it on my resume, I'm making double my salary next year. If I can find a simple ruleset or linear regression that performs 90%+ as well as the XGBoost and put it in production then nobody cares even though it feels like distilling the complex down to the simple is really where the value is.
The thing is that anymore anything short of a deep net is equally difficult to implement as a linear regression, if not easier due to nulls and categoricals.
Also, modern tooling makes a lot of these models more than explainable enough for a lot of cases… 10% is a lot
Just as an FWIW, I've been interviewing data people for about a decade now, and I would definitely be far more impressed with the simpler approach, but I realise that people like me are a minority in the field.
Also agree about the simple tools but it's really hard from a career perspective. If I deploy XGBoost in production and put it on my resume, I'm making double my salary next year. If I can find a simple ruleset or linear regression that performs 90%+ as well as the XGBoost and put it in production then nobody cares even though it feels like distilling the complex down to the simple is really where the value is.