Sponsored by: Zingage
We’re turning home care into software.
$300B is authorized each year for care at home. $90B never gets delivered. Zingage is the AI operator closing that gap — an always-on back office running 10,200 visits and 11,000 conversations a day for home care agencies nationwide.
Welcome back, everyone —
Here’s what we’re covering this week:
AI outperforms doctors, again
AI catches cancer years earlier
AI-first providers win the future
24 new tools/partnerships, 11 funding updates, new AI jobs & link-worthy content
Read time: 5 minutes
Our Picks ✨
Highlights if you’ve only got 2 minutes…
1/
AI outperforms doctors, again
A new Harvard study that went viral this past week is adding fuel to the AI in medicine debate. Researchers found that an OpenAI model diagnosed real ER patients more accurately than physicians, even in the chaotic early stages of care. On 76 real cases from a Boston hospital, the AI got to the correct or close diagnosis 67.1% of the time, compared to 55.3% and 50.0% for two attending doctors.
To add to it the model used was o1 preview, released back in 2024. o1 has already been surpassed by multiple new models, from o3 through GPT-5.5. On benchmarks like HealthBench, newer models are showing roughly 35% better performance. Despite being “old” by AI standards, it still outperformed clinicians, and reviewers could not reliably distinguish between AI and human diagnoses.
The learning here is less about AI versus doctors and more about direction. The capability is here and improving quickly. The real challenge now is integration, figuring out how these systems actually fit into clinical workflows. (link)(linkedin)
2/
AI catches cancer years earlier
A new AI model from Mayo Clinic could change how one of the deadliest cancers is detected. The system, called REDMOD, can identify signs of pancreatic cancer on routine CT scans up to three years before a typical diagnosis, way before tumors are visible to the human eye.
In a study of ~2,000 scans that were originally read as normal, REDMOD flagged 73% of cases that were later diagnosed with cancer. Two years ahead of diagnosis, it identified roughly 3x more early cancers than experienced radiologists. The model works by analyzing subtle tissue patterns, texture, and structural signals that clinicians cannot see.
Pancreatic cancer has a five year survival rate below 15%, largely due to late detection so the impact here could be significant. There’s no greater mission than using tech to help people live healthier lives. This gets us hyped. (link)(twitter)
3/
AI-first providers win the future
BCG makes a pretty direct case that healthcare is at a breaking point and AI is no longer optional. Demand for healthcare is expected to outpace supply by about 14% through 2036, while labor costs keep rising faster than reimbursement, putting pressure on already thin margins. At the same time, healthcare spending is growing ~ 3x faster than GDP, making the current model unsustainable.
Despite the momentum, most organizations are still stuck in pilot mode. Nearly 78% of AI efforts fail to scale, largely because systems try to layer AI onto existing workflows instead of rethinking how work gets done. That’s the unlock. BCG estimates 70% of the value comes from operational and organizational change, not the technology itself.
Where it is working, AI is already improving scheduling, documentation, triage, and billing. The real question now is who is actually rebuilding around it. (link)(linkedin)
TOGETHER WITH SUPERDIAL
Agentic voice AI for healthcare RCM. SuperDial navigates IVRs, waits on hold, and completes payer workflows autonomously. 👉️ Click here to learn more!
Tools & Partnerships 🔧
Latest on business, consumer, and clinical healthcare AI tools and partnerships…
TOOLS
Cursor launches Life Sciences offering: The company is expanding into healthcare and biotech, aiming to accelerate drug discovery, diagnosis, and research with a model-agnostic AI platform. (link)
Google DeepMind advances AI co-clinician: Multimodal system showed near-zero critical errors and matched or outperformed physicians in many tasks, but still lagged in identifying red flags. (link)
Hugging Face hub aggregates science datasets: New initiative centralizes massive open datasets—from genomics to medical QA—to accelerate research and model development. (link)
OpenEvidence pulls out of Europe: The clinical AI app exited the EU and UK citing regulatory uncertainty around the AI Act, raising concerns other companies may follow. (link)
CCS scales agentic AI for chronic care: Its CeeCee platform automates patient interactions and supply workflows, aiming to improve access and operational efficiency. (link)
Claude shows strong bioinformatics research skills: On a new BioMysteryBench benchmark, AI matched expert-level performance and even solved some problems humans couldn’t, highlighting rapid progress in scientific reasoning. (link)
JAMA proposes new AI regulation model: Authors argue autonomous clinical AI should be regulated more like physicians, with licensure and ongoing oversight. (link)
FDA pilots real-time clinical trials: New trials allow the FDA to monitor safety signals during studies, aiming to improve oversight and speed up research. (link)
AI uses facial photos to predict cancer survival: A Mass General Brigham model tracks changes across patient photos over time to better forecast outcomes and guide care. (link)
AI predicts lung surgery risk more accurately: A new model combining imaging and clinical data outperformed surgeons in identifying postoperative complications. (link)
Goodfire launches AI “debugging” platform: Silico lets researchers inspect and intervene in model internals, helping uncover mechanisms behind predictions, already enabling advances in genomics and Alzheimer’s research. (link)
UnityAI optimizes staffing with agentic AI: StaffOps uses real-time EHR data to match workforce scheduling with patient demand and reduce inefficiencies. (link)
Arintra adds AI for documentation improvement: New features identify coding and documentation gaps, boosting revenue and reducing denials. (link)
Freed launches AI front desk for clinics: New receptionist tool handles calls and routes structured requests to staff automatically. (link)
Atropos releases real-world evidence platform: Alexandria provides a large dataset for clinical decision support and AI model development. (link)
HOPPR turns imaging into structured reports: New AI model converts chest X-rays into narrative summaries for radiology workflows. (link)
Eleos expands AI across behavioral health ops: New agents support eligibility, documentation, revenue cycle, and compliance workflows. (link)
Open Wearables simplifies health data access: New open-source platform unifies wearable data across devices and enables LLMs to reason over it with transparent, customizable scoring. (link)
PARTNERSHIPS
Beth Israel Lahey Health + Heidi: Beth Israel Lahey Health is rolling out Heidi’s ambient AI platform across 6,000 clinicians to improve documentation and reduce after-hours work. (link)
UChicago Medicine + Artisight: UChicago Medicine signed an enterprise agreement with Artisight to deploy its Smart Hospital Platform across care settings, including patient rooms, ORs, and a new cancer facility. (link)
Viz.ai + National Rural Health Association: Viz.ai partnered with the National Rural Health Association to provide rural health systems with AI education, toolkits, and implementation guidance. (link)
Graybill Medical Group + Third Way Health: Graybill partnered with Third Way Health to deploy an AI-enabled operations platform, reducing front-office costs by 50% and improving patient access. (link)
Credo AI + Coalition for Health AI: Credo AI joined the Coalition for Health AI to advance governance frameworks for safe and scalable AI deployment in healthcare. (link)
WHOOP + U.S. Navy: WHOOP partnered with MIT Lincoln Laboratory to deploy wearable technology for monitoring fatigue and readiness in naval personnel. (link)
Deal Desk 💰
Spotlight on latest capital raises, M&A, and investments…
FUNDING
Aidoc, an Israeli medical imaging platform, raised $150M in Series E funding. Goldman Sachs Alternatives led, joined by General Catalyst, SoftBank Investment Advisors, and NVentures. (link)
Iterative Health, a healthcare technology and services company powering the acceleration of clinical research, raised $77M in a Series C led by Intrepid Growth Partners and Google Ventures. (link)
Fathom Therapeutics, a NYC-based quantum and AI drug design company, raised $47M in Series A funding. Sutter Hill Ventures led. (link)
Enzo Health, a Lehi, Utah-based post-acute care startup, raised $20M in Series A funding. N47 led. (link)
Hypervision, a London-based developer of hyperspectral imaging for surgery, raised £17M in Series A funding. Heal Capital led. (link)
Illuminant Surgical, a precision access and visualization company to navigate the human body, raised $8.4M in seed funding led by Wing 2 Wing Ventures. (link)
Gravity Rail, AI-powered engagement workflows for healthcare, raised a $2.75M seed investment from Redesign Health. (link)
Betterness, a Miami, Fla.-based developer of autonomous AI agents for personalized wellness, raised $2.5M in seed funding. (link)
AIRA Health, a Boston-based developer of decision infrastructure for modern clinical development, raised $2M in pre-seed funding. (link)
Thea, building AI-Doctors focused on the aesthetic industry, launched and is backed by General Catalyst, Stellation, & other incredible investors. (link)
GRANTS
Chan Zuckerberg Biohub commits $500M to AI biology: The nonprofit is investing in advanced AI models to simulate the human body and accelerate biomedical research. (link)
as of 5/3/26
Other Relevant News 🔍
News, podcasts, blogs, tweets, resources, etc…
AI Job Opportunities 💼
Explore our AI Job Board or contact us to feature roles in our newsletter…
Visuals of the Week 📸
Funny memes, cool pics, and interesting data from around the web…
That’s it for this week friends! Back to reading — I’ll see you next week.
Stay classy,
PS. I write this newsletter for you. So if you have any suggestions or questions, feel free to reply to this email and let me know
How was this week's newsletter? Tap your choice below👇
