Anthropic just handed drug labs a model that designs molecules ten times faster.

Run time: 13:30

In today's episode:

1. Ambient AI scribe cuts clinician burnout in randomized trial
2. Philips survey: AI saves sixteen workdays, training lags
3. AIRS SwiftMR wins wider FDA MRI clearance
4. Mount Sinai launches AI cancer trial-matching tool
5. Nature Medicine: AI diagnosis accuracy swings fourfold
6. Topol says every mammogram should use AI
7. Anthropic ships Fable 5, its strongest public model
8. Claude Code adds parallel-subagent dynamic workflows
9. GPT-5.6 and Gemini 3.5 Pro expected this month

TL;DR:

• A pragmatic RCT of an ambient AI scribe cut clinician exhaustion and documentation time without degrading note quality or billing accuracy — the kind of evidence the field keeps promising and rarely delivers.
• Anthropic released Claude Fable 5 (and the unlocked Mythos 5 for vetted bio/security researchers); scientists preferred Mythos 5's molecular-biology hypotheses over Opus-class output ~80% of the time, with a documented ~10x drug-design speedup.
• Philips' Future Health Index says AI already saves clinicians 16 working days a year — while 70% of those same clinicians call their AI training inadequate or nonexistent.

Sources cited:

• NEJM AI
• GlobeNewswire
• Diagnostic Imaging
• Mount Sinai
• Nature Medicine
• Ground Truths
• Anthropic
• Releasebot
• Essa Mamdani
• Primer

Subscribe: YouTube

medAI Times is for educational and informational purposes only. The content does not constitute medical advice, diagnosis, treatment recommendation, or professional clinical guidance. Consult qualified healthcare professionals and refer to official sources before making clinical, research, regulatory, or business decisions.