Yeah their built-in nuclear data, for instance, has never had sufficient citation info to use it for actual research, and it's unclear who would play with such

Yeah their built-in nuclear data, for instance, has never had sufficient citation info to use it for actual research, and it's unclear who would play with such information without having any possible interest in publishing. Mathematica seems to be a grab bag of whatever excites Wolfram recently. It was heavy on cellular automata a few years ago and now it has llm stuff. It is a nice environment for playing with computation and doing back of the envelope work, but in the end i always write python programs instead because it's just easier to work with real data that way and the mathematica licensing is a pain.