How a researcher uses Codex and ChatGPT to search for new antimicrobial molecules
- ID
- 23218
- Status
- summarized
- Published
- 11 Sep 2026, 12:00 AM
- Fetched
- 11 Sep 2026, 12:25 AM
- Provider
- OpenAI News
- Category
- ai-labs
- Original URL
- https://openai.com/index/using-codex-chatgpt-to-search-for-new-antimicrobials
- Source URL
- https://openai.com/news/rss.xml
Summary
- Score
- 3.5
- Created
- 11 Sep 2026, 12:28 AM
- Tags
- Audience
- ai_ml_learnersai_agent_users
What happened
OpenAI profiles César de la Fuente's bioengineering lab, which uses deep-learning models to search genome and protein datasets for antimicrobial candidates, compressing an initial discovery phase from years to hours. The lab also uses ChatGPT and Codex for hypothesis brainstorming, code writing, dataset processing, and cross-disciplinary idea connection. The piece frames antimicrobial resistance as a growing crisis with ~5 million associated deaths in 2021, projected to roughly double by 2050.
Why it matters
This is a vendor-published case study with minimal technical detail—no model architecture, no specific molecules discovered, no benchmark results. Builders should not change tooling decisions based on it. The only transferable idea is the workflow pattern of using LLMs alongside custom ML models for literature-style hypothesis generation and code scaffolding, which most AI/ML practitioners already do.
Discussion angle
Where is the line between a genuinely useful AI-in-science workflow write-up and a vendor promo? What specific technical details would have made this actionable for builders versus what we got?