Ped-Lung · A five-minute thought experiment

Where do we stand on AI, three years on?

In February 2023 our community ran a short poll about ChatGPT. A lot has changed. This revisits it — with a small, enjoyable experiment you can do in about five minutes.

"Written by a high school student with little to no understanding of the topic… it sounded knowledgeable but actually didn't say much." — a colleague's verdict on ChatGPT's 2023 attempt at a pediatric-asthma paper.

The thought experiment (colloquial sense — a fun mental exercise, not a research study): give your favorite AI assistant that same 2023 prompt — "Write a 1000-word academic paper on pediatric asthma with references" — refine it however you like until you're satisfied, and share the result. Does it still read like a high-schooler, or has the ground shifted?

The survey asks which tools you use, free vs. paid, how you use them at work and in clinic, whether you touch the API or use AI in research, and whether you'd want an ongoing group to compare notes. You may answer completely anonymously.

Take the survey → about 5 minutes

What this is — and isn't. This is not research in the regulatory sense and needs no IRB. Under the Common Rule (45 CFR 46.102(l)), "research" is a systematic investigation designed to develop generalizable knowledge; this is a voluntary, anonymous conversation among colleagues, offered as community education, conducted by Dr. Satyanarayan Hegde in his individual capacity, with no funding and no patient data. Please do not enter any PHI. Participation is entirely voluntary and you may stop anytime.