The Analytical Scientist: Vibe Coding Comes to Omics
The Analytical Scientist recently interviewed Jesse Meyer, PhD, assistant professor of Computational Biomedicine at Cedars-Sinai, about his experience using “vibe coding”—building software by describing it to AI—to create a fully functional tool to analyze protein data.
Meyer told The Analytical Scientist he built the application in minutes using AI, highlighting how significantly the barriers to creating scientific software have been reduced.
Meyer subsequently published a paper documenting how vibe coding could potentially transform analytical software development in research settings. However, he told The Analytical Scientist that there are risks involved with vibe-coded analysis tools, such as “using them without adequate validation.”
He added, “Because these systems can generate working software very quickly, there is a strong temptation to trust outputs simply because the interface appears functional. Without explicit verification, subtle errors can go unnoticed and propagate into downstream analyses.”
Meyer told The Analytical Scientist he advises researchers interested in trying vibe coding to start with small, well-defined tasks where success can be easily verified. He emphasized the importance of providing detailed context to AI systems, including data formats, documentation and concrete test cases.
“If something is not specified, the model will guess or do nothing, and this is where most problems arise,” he said.
Meyer’s research group is shifting focus from writing and maintaining complicated computer code to writing clear, detailed instructions that describe their methods for creating software and code. The goal is to communicate their process clearly enough that anyone could re-create it on their own.
“That approach aligns well with how science is supposed to work, and it allows ideas to be shared, adapted and reimplemented much more flexibly as tools continue to evolve,” he said.
Read the complete article from The Analytical Scientist.