Distinguished Engineer at LinkedIn and Professor of Statistics at Harvard University (on Leave)
I asked o3 (deep research) to create a statistical scenario that it doesn't know how to answer. Perhaps one can use this system prompt to see how models evolve similar to "draw a unicorn using latex". TLDR: it picked causal inference. https://lnkd.in/dPiGSC2g
Richard Hahn
2mo
Interesting...that "deep dive" is sort of all over the place. Much like LinkedIn posts, it's a grab-bag of different controversies in statistics. Which I guess makes sense. As for the causal inference angle, what is notable is arguably causal inference isn't really a *statistical* problem, per se. I'd have been more impressed if it had said "fiducial inference" :-)
Draw an SVG of a pelican riding a bicycle doing causal inference :-)
Data Scientist | Machine Learning Engineer | Master's degree in Engineering
2moI was hoping to see more of a scenario illustrating a new "paradigm" or solution, rather than the typical limitations of a statistical exercise in real life. Maybe I’m misunderstanding, but it seems like there’s a mix-up between the usual issues in causal analysis and the kind of real-life scenario that might call for a new statistical model. What do you think?