👀Using an autonomous agent based on o3-mini and GPT-4.1, a team from Harvard, MIT & other institutions reproduced and updated an entire issue of Cochrane Reviews in two days… saving 12 person-years of work!
The AI reviews captured more papers & were more accurate than humans.
Ophthalmology notes are full of acronyms - “CME, PPV, NAION."
In our @MayoClinic randomized QI study, we added plain language summaries and tested them in real care settings. 85% of non-eye clinicians preferred Opthalmology notes with an AI-generated summary.
#MedTwitter#AI#LLM
Integrating large language model-generated plain language summaries into ophthalmology notes can improve interdisciplinary communication, although ongoing review and safety monitoring are necessary for successful clinical implementation. https://t.co/WCqt3kIiIU
Still, AI isn’t foolproof. Even after physician review, 15% of summaries had errors on blinded overread. The key: 0% were deemed harmful. AI can support communication—but only with physician oversight. It’s a tool, not a replacement.
While most ophthalmology journals have adopted guidelines on generative artificial intelligence (GenAI), this study found variability in their scope and comprehensiveness. https://t.co/tmvZ1TXXX0
Review by @PrashantDTailor@HaleyDSouzaMD@MRStarrMD discusses the evolving role of large language models in ophthalmology, focusing on current applications and future potential in enhancing ophthalmic care. In @CO_Ophthalmo https://t.co/3skycOinHu
Had a great time serving on the faculty at the 18th Annual Heed Retreat this week. Enjoyed discussing careers in academic #ophthalmology with a very accomplished group of residents and sharing my insights on career development.
We just released Mixtral-8x22B-v0.1 and Mixtral-8x22B-Instruct-v0.1:
- Free to use under Apache 2.0 license
- Outperforms all open models
- Native function calling
- Masters English, French, Italian, German and Spanish.
- Seq_len = 64K
https://t.co/SCG8s06Dbl
Randomized, masked, multicenter study found that #LLM responses to common #retina questions were comparable to experts in terms of quality, empathy & safety metrics. @MRStarrMD@mayocliniceye@prashantdtailor https://t.co/ywuusFRSJ0
Totally impressed by the Google Scholar PDF Reader Browser Extension! It's completely transformed how I check references on the fly.📚 #GoogleScholar
🔗 https://t.co/cECgqQyYkp
A packed morning session on day 2 of Mayo ID Updates 2024 with fellow Dr. @RyanKhodadadiMD presenting unusual and difficult to manage infections to faculty Drs. Steven Holland and Avni Joshi
Diagnoses:
1. Chronic granulomatous disease
2. Anti-interferon gamma autoantibodies