Member-only story
I am so over data science …
but ChatGPT would be good at Jeopardy.
I remember when the words data science became part of my language. I started working on a massive project with IBM about a decade ago.
IBM had just come off their Jeopardy win. People were impressed. Watching a computer beat people at such a nuanced game was an entertaining milestone.
People that built algorithms for their day-to-day knew it was a big deal. IBM had built a robust and complex algorithm that could go into the wild and not embarrass anyone.
I spent years on that project with IBM. Our goal was to apply the same technology to the biomedical literature. Early on, aspects of the projects were funded by DARPA. Over a three-year period, we published a few papers and created a product that sold to Pharma companies for millions of dollars. I still remember my IBM colleagues fondly.
My IBM colleagues told me it was, for a moment, a flagship project for IBM. Apparently, Warren Buffet liked it in the Boardroom.
The problem is that I never felt like it worked. The product could be helpful but was limited. It could organize information, but the ability to ask a question was far away. It was too hard of a problem at that time.
