What you get every issue:

Three curated healthcare AI stories with my take. I'll list companies doing the work, AI Tools, skills and techniques worth knowing, and opportunities worth pursuing—no hype just signal.

Three things worth your time this week:

  1. AMA SURVEY: 81% OF PHYSICIANS NOW USE AI — BUT 88% FEAR LOSING THEIR SKILLS. Physician AI adoption doubled from 38% to 81% in three years. Average use cases per doctor jumped from 1.1 to 2.3. The headline: "Physician AI adoption skyrockets." The reality: 88% are worried about skill atrophy. And the fear is strongest among younger physicians — the ones who should be most comfortable with it. My take: Nobody's designing for the dependency problem. If your AI makes doctors faster but less capable, you haven't solved burnout — you've traded it for fragility. Read it: [AMA 2026 Physician Survey]
  2. GOOGLE KILLS CROWDSOURCED AI HEALTH FEATURE AFTER SAFETY BACKLASH. Google removed "What People Suggest" — an AI feature that surfaced amateur medical advice from Reddit and forums. Came after an investigation found AI Overviews putting lives at risk. Why it matters: Even Google can't shortcut clinical trust with crowdsourced content. The pattern: Consumer health AI is retreating. Enterprise health AI is restructuring. The middle ground is a graveyard. My take: They launched a feature that let strangers give medical advice through AI. The fact it shipped at all tells you everything about the gap between tech ambition and clinical reality. Full story: [The Guardian / Multiple Sources]
  3. VERILY RAISES $300M AND BREAKS FREE FROM ALPHABET Verily becomes Verily Health Inc. Alphabet drops controlling stake. Samsung Galaxy Watch and Salesforce Agentforce partnerships were signed within weeks of independence. But here's the nuance: This isn't just a funding round. It's a structural play for vendor neutrality. No health system wanted to share data with a Google subsidiary. Why this matters for builders: Independence unlocks enterprise trust. If you're a health tech subsidiary of Big Tech, this is your playbook. If you're competing with one, your window just got smaller. Source: [Verily / MobiHealthNews]

COMPANIES DOING THE WORK Verily Health — Precision health AI platform, newly independent from Alphabet, building ecosystem with Samsung and Salesforce. Mindgard — AI security firm that exploited Doctronic's AI prescriber to triple opioid doses. Proving that red-teaming clinical AI isn't optional. Mount Sinai Research Team — Published the study showing multi-agent AI holds accuracy at clinical scale, where single agents crater to 16%.

AI TOOLS WORTH KNOWING

Amazon Connect Health — AWS — Agentic AI for healthcare admin: scheduling, verification, documentation, coding. First to build evidence mapping that links every AI output to its source record. UC San Diego saving 630 hrs/week. The trust benchmark just got set.

Dragon Copilot + Marketplace — Microsoft — 100K clinicians on ambient AI documentation. Now an open marketplace for third-party clinical AI apps. If you're building a health AI product, this is a distribution channel.

Elation Health AI Billing — Elation Health — Touchless claims processing for primary care. 72% of eligible claims auto-generated, 95%+ submitted without manual intervention. Targeting sub-5% denial rates. This is revenue cycle AI in production, not in a pitch deck.

Thyme Care Acuity Engine — Thyme Care — AI that synthesizes claims, clinical, and SDOH data to dynamically prioritize highest-risk oncology patients for human social worker outreach. 30% reduction in ED/inpatient events. AI for triage, not diagnosis.

WORK & PARTNERSHIP OPPORTUNITIES VP Product, Precision Health Platform — Verily Health — $300M raise plus Alphabet spinout means massive product hiring. Clinical background preferred. Director of Clinical AI Validation — Multiple companies post-HIMSS — Validation gap exposed across AI agents from Epic, Google, Microsoft, and Oracle. Rare skillset, high demand. Head of AI Safety, Healthcare — Open across industry — Doctronic exploit proved AI prescribers can be manipulated. Every serious health AI company needs this role yesterday.

WHAT I'M WORKING ON: Next week: "The Validation Gap" — why every major HIMSS launch this year is missing the same thing, and what it costs when production hits.

— Victor

Reply to this email if you want to discuss any of these. I read every message.

Yuuki Edge • Healthcare AI Intelligence • yuukiedge.com