I read that Apple is falling behind in AI research because it’s cult of secrecy makes it difficult to attract top-level researchers.
This resonated with me, because when I worked in finance in the late 80s and the 90s, I often felt that some of the people I worked for didn’t understand the motivation of researchers.
The group I ran did research on volatility and built options models. We began by publishing nothing, because the people that managed us had the gut-level feeling that giving away any ideas was not a good idea at all. But I came from a physics research background, read academic finance papers, and felt as though I was part of a community of researchers to which I wanted to contribute. There was a selfish facet to this — all researchers, like writers, want acknowledgment — but it was largely good selfishness, the Adam Smith kind. And so I pushed to publish. And slowly, the people I worked for, especially the salespeople, began to realize that publishing was often a good business strategy too.
If you produced relevant research, and customers or clients (should I call them guests as Duane Reade and Starbucks now call me?) knew about it, they were more likely to bring their business to you and to seek your insights and advice. And advice is ultimately what financial customers need — financial securities have many confusing parameters and you need a model to project these multidimensional attributes onto a one-dimensional scale that ranks securities by value.
Some managers project the entire world onto a one-dimensional scale of compensation and its correlate, power. But researchers want peer recognition too, and perhaps need it to do their best work.