Wednesday, 31 January 2024: 9:15 AM
338 (The Baltimore Convention Center)
More and more, the environmental sciences are moving toward solutions that require an expertise in programming - from analysis of data, to machine learning, to basic data access. For a scientist new to programming, or even those experienced, the process of writing and testing this code can take hours. Not to mention the time spent debugging mysterious error codes!
However, this workflow is about to change - with AI. Specifically, large language models (LLMs), which have shown an incredible ability to write, read, and even debug code. In the past few months, there has been a large shift toward a new workflow: AI-Assisted Programming. This workflow allows you to spend more time planning the purpose of your code while letting a LLM go through the drudgery of debugging. This presentation will walk through an AI-assisted programming demo given for oceanographers at OceanHackWeek 2023 and overviews the theory behind AI-Assisted Programming, important considerations when working with LLMs, as well as recent success stories of AI-Assisted Programming in action.

