Journal

PLAB 1 fell from 74.1% in 2021 to 62.2% in 2025, with PLAB 2 at 59.8%; what the official GMC data does show, the explanations it cannot support, and what candidates should actually...

70% in one question bank is not 70% in another, and neither converts into an exam score; difficulty, repetition, adaptivity and equating explained, plus a readiness dashboard that ...

From May 2026, Step 2 CK runs as sixteen 30-minute blocks and Step 1 as fourteen; total day length and item counts were not doubled, the questions were redistributed, and pacing is...

From the 2026/1 diet the MRCP Part 1 pass mark is 450, an equated scaled score, not 45% and not a fixed number of correct answers; what that means for revision and question-bank pe...

A founder's account: why a TEDx talk about undifferentiated patients became an AI tutor that refuses to just answer, and why medical education needs fewer answers and better questi...

Medical education is organised by chapter; patients are not, and that gap is where clinical reasoning lives. Why generative AI, used Socratically, could finally close it.

Three genuinely different ways AI now teaches clinical reasoning: Quesmed's OSCE simulator, Geeky Medics' virtual patient ecosystem, and iatroX's Socratic interrogation, and why th...

Geeky Medics has quietly built one of the most ambitious AI stacks in medical education, tutor, virtual patients, AI examiners; here is how it compares with iatroX's reasoning-firs...

AMBOSS calls its AI Mode Learning a study copilot; iatroX builds the copilot around learner performance: two credible visions of the same product category, compared honestly.

The static question bank is quietly becoming obsolete as a category: from analytics to adaptive learning to generative AI to full orchestration, how Pastest, PassMedicine, AMBOSS, ...

Pastest now offers on-demand AI help while you answer questions; iatroX makes the AI interrogate your reasoning: two philosophies of AI medical revision, compared on educational me...

Medly and iatroX serve different learners and represent the same shift: vertical, curriculum-aware AI tutoring; here is how the model changes when the subject becomes medicine.