Coding agents provide lightning fast data visualization tools

In this completely wild interview Peter Thiel talks about his views on stagnation and how science and society in general is not progressing forward. He says that in some ways specialization has made it so that moon shot projects are not happening, but also, moon shot projects are simply not happening. He makes some sense about this in my opinion. Specialization in science makes it such that a collection of peer reviewers (typically 3) can no longer adequately review papers because a paper includes too many different topics. For example, you might want someone with a background in AI that includes LLMs and computer vision, someone with a background in geochemistry, and someone with a background in geophysics to review t his paper I wrote about a data pipeline for ingesting multi-modal borehole data. But just asking for adequate reviewers for the AI section is perhaps a tall order when there are 100 AI/ML papers posted on Arxiv a day.

Interactive data visualization is a a key skill that has, for a long time, been locked up by front end web development tools ( d3.js for example) that most scientists get no exposure to during their PhD training.

Leave a Comment