Released on June 22, 2023
Back to episode listEpisode Title: AI Unleashed: Navigating the Opportunities and Challenges of AI in Microbial Bioinformatics
Hosts: Lee, Nabil, and Andrew
In this episode of the Micro Binfie Podcast, titled "AI Unleashed: Navigating the Opportunities and Challenges of AI in Microbial Bioinformatics," hosts Lee, Nabil, and Andrew explore the implications of generative predictive text AI tools, notably GPT, on microbial bioinformatics.
Application of AI Tools: The hosts discuss various applications of AI in their work, from generating boilerplate programs and drafting documents to summarizing large datasets. Andrew highlights his experience using GPT in coding through VS Code and GitHub Copilot, where GPT can generate nearly 90% of the necessary code based on a brief task description, significantly accelerating his work.
Experience with GPT: Andrew discusses using GPT to clarify lines of code and how AI was utilized to draft a paper on the ethical implications of AI in microbial genomics research during a recent hackathon.
Nabil's Insight: Nabil shares his experience with GPT in standardizing date formats in tables and summarizing paper abstracts. While GPT is typically accurate for simple tasks, he warns of the potential for erroneous answers in more complex scenarios, as he experienced when using GPT to find a route in a video game.
Challenges of GPT: Andrew outlines a script created during a hackathon for producing podcast episodes reviewing math tools and discusses challenges like GPT providing incorrect factual information.
Future Prospects and Challenges: Looking ahead, Andrew envisions a future filled with GPT-generated content that may often be incorrect, heightening the challenge of distinguishing between real and false information. The potential benefits for individuals with visual impairments are acknowledged, though current solutions are not yet perfect.
Live Experiment: Nabil and Andrew conduct a live experiment, using a Perl script and GPT-4 to translate it into Python and back. The AI tool demonstrates proficiency, intelligently using popular libraries like BioPython, though it includes a disclaimer about potential inaccuracies.
Optimization and Language Learning: They discuss the possibility of using AI to optimize code, akin to minifying JavaScript, and iterating through multiple languages to assess outcomes. Nabil tasks AI with writing a Python script that translates DNA into protein, which is then translated into Rust.
AI in Complex Tasks: Andrew shares his experience of using AI to generate a Python class that compares two spreadsheets using pandas, showcasing AI's capability in comprehending and executing complex tasks.
The episode underscores the power and potential of AI in coding while emphasizing the necessity of human oversight to ensure quality and accuracy in AI-generated content. It envisions a future where AI tools could revolutionize aspects of programming, introducing new efficiencies and methods of working, despite current limitations.