AI in Healthcare

Today's Stories

1. 1

   Asus VP: hospital AI barrier isn't models, it's data readiness

   Asus vice president Kuo-Hao Huang, of the Asus AI Research Center, said hospitals' biggest obstacle to adopting AI is whether their data, workflows, and system architectures are ready, not the AI models. Asus is pushing xHIS as a digital backbone for AI-ready healthcare, pairing AI, hardware, software, and cloud; Huang's view shifts the readiness test from model choice to hospital IT foundations.

   The outcome hinges on whether hospitals modernize data and workflows to match Asus's stack. Watch xHIS adoption as the signal of whether that shift happens.

2. 2

   CDD Vault, Novartis urge data foundation before drug discovery AI

   In an Emerj series sponsored by CDD Vault, Novartis's Xiong Liu and CDD Vault's Barry Bunin urged R&D leaders to fix fragmented data and annotation before scaling AI. The guests said drug discovery typically takes 10 to 15 years and roughly 90 percent of candidates fail, and that siloed, poorly annotated data blocks model training and reproducibility.

   Whether teams treat metadata consistency as a gate and can show both secure data and value sizing; proposals that answer both tend to move faster, according to the guests.

3. 3

   Thermo Fisher's Sarin: diagnostics partners must show value, not just sell

   At ADLM, Thermo Fisher Scientific's Puneet Sarin said global demand for advanced specialty diagnostics remains strong across North America, Europe, Latin America, Asia Pacific, the Middle East, and India, and that customers are rethinking how they view partners. Sarin said the notion of partnership is moving beyond instruments and reagents toward companies that demonstrate clinical and operational value creation and can operate at scale, as geopolitical factors, inflation, and reimbursements weigh on customers.

   Sarin describes the clinical lab of the future as a convergence of automation, artificial intelligence, and new business models, and says Thermo Fisher aims to be the trusted partner enabling that transition; the test is whether it can support customers moving to panel-based diagnosis and less invasive monitoring.

4. 4

   Doctor scolds patient who came after ChatGPT consult

   An office worker who felt chest tightness asked ChatGPT about his symptoms, was told heart or lung problems could not be ruled out, and went to a hospital. The doctor sighed and asked why he came just because AI said so. The tests found no serious abnormality and the doctor said there was little to worry about, yet the visit still drew a rebuke. The poster said he never used AI to overrule the doctor's diagnosis or misuse medicine.

   The dispute hinges on whether AI symptom-checking should count as a reason to seek care, or as needless anxiety. Commenters split, with some saying many patients blindly trust AI and others saying AI cannot make a medical judgment.

5. 5

   Apple's SimpleDesign hits competitive protein design with single-stage training

   Apple researchers introduced SimpleDesign, a single-stage end-to-end model that trains directly on over 2M sequence-structure pairs, combining discrete cross-entropy for sequences with a regression objective for structures. Existing co-design models rely on a two-stage process, first training autoencoders then training a generative model in latent space; SimpleDesign skips this by training in data space, achieving competitive performance across benchmarks.

   The test is whether this single-stage approach holds up for larger proteins or other multi-modal biological tasks, which the body does not address. The model uses Transformer-based multimodal backbones with modality-specific processing and global self-attention.

6. 6

   Incredible Health's AI agent Lyn cuts hiring time 30%

   Iman Abuzeid, co-founder and CEO of Incredible Health, said the company's AI agent Lyn has reduced hiring time by 30%, with three-quarters of interviews now within 24 hours of applying. Interviews that used to take up to two weeks now mostly happen within a day, and about 40% occur at night or on weekends when recruiters are less available. Employers apply to workers, reversing the usual process.

   The test is whether the time savings hold as more employers adopt the platform, which is free for healthcare professionals and paid for by employer subscriptions. Watch the roughly $97 million raised across seed, Series A and Series B.