20-Minute Fine-Tuned 2015 CNN Beats GPT-6 and Claude at Skin Lesion Diagnosis

A dermatologist found that a 2015 Inception v3 CNN fine-tuned in 20 minutes achieved 62% accuracy on 200 biopsy-confirmed skin lesions, beating GPT-6 and Claude. Laurence Moroney cited the test in a blog arguing narrow models still matter.

2026-10-26 ~ 2026-10-26 · 3 related posts

• Dermatologist: a 10-year-old CNN beats frontier LLMs at skin cancer diagnosis, 62% vs 42% — RazRazcle · 2026-10-26
• Fine-tuned 2015 CNN beats GPT-6 and Claude on skin lesion diagnosis, 62% vs 42% — lmoroney · 2026-10-26
• Laurence Moroney: why narrow fine-tuned models still beat frontier LLMs, with a selection rule — lmoroney · 2026-10-26