This book is the second volume from a series based on thoughtful, conversational exchanges between the author, an experienced infectologist from Brazil, and his personal AI. Written in an accessible style for both medical professionals and lay readers, it draws from the author's real clinical cases, publications, and reflections on medicine, science, education, and public health. The book opens with a discussion of a popular 2025 YouTube video revisiting Richard Feynman's 1950s critique of Brazilian education: students excelled at memorization but struggled with genuine understanding and application. The author and AI explore how this "knowing by heart but not understanding" issue persists in Brazilian schools (evidenced by ongoing low PISA scores) despite gains in access, and they link it to medical practice, where rote protocols can miss nuanced diagnoses.A key clinical thread begins with a striking emergency room case: a woman with unexplained right-sided chest pain, normal imaging, but a classic dermatomal pain pattern described by hand gesture. The author diagnosed herpes zoster sine herpete (shingles without rash), treated it successfully with acyclovir and betamethasone, and contrasts this intuitive, pathophysiology-driven reasoning with the superficial approach Feynman criticized. Other cases include:
- A patient with chronic kidney disease, anemia, unusual thrombocytosis (~700,000/mm³), lower back pain, and furosemide use. Initial suspicion of multiple myeloma was reasonable but unconfirmed; the patient died rapidly from superimposed infection. The author later published a scoring system to prompt consideration of myeloma in similar presentations.
- A 2012–2013 case of immunovirological discordance in early ART (CD4 rise but persistent/high viral load), later published, revealing transmitted resistance.
- Reflections on older work, including a 2010 review of HIV superinfection (controversial at the time for social/sexual implications) and a 2011 report of post-malaria neurological syndrome (PMNS), an immune-mediated condition responsive to steroids.
The book closes with extended discussions on COVID-19: family clusters with multiple ARDS/deaths, genetic predispositions (including HLA alleles like C*04:01 linked to cytokine storm risk), population frequencies of risk alleles, critiques of prolonged general lockdowns (advocating focused protection of vulnerable groups), and perceptions that some indefinite sanitary measures were driven by liability avoidance rather than pure evidence. Overall, the work celebrates deep clinical reasoning, lifelong learning, the value of publishing real cases, and the evolving interface between medicine, genetics, immunology, and public policy — all framed through candid, encouraging dialogue with AI.
