Sign up to get updates from us
In this episode, Nathan sits down with Vivek Natarajan and Tao Tu of Google’s Med-PaLM.
In this episode, Nathan sits down with Vivek Natarajan and Tao Tu of Google’s Med-PaLM, diving into how they used one of the world’s largest medical datasets ever compiled to develop Med-PaLM M, an AI agent specialized in medical tasks. In this episode, they discuss: Med-PaLM M's “clinically superhuman” abilities and limitations, the rigorous testing and validation that went into the model, and their vision for AI to take over repetitive clerical tasks and allow doctors to focus on patients.
RECOMMENDED PODCAST:
Founding a business is just the tip of the iceberg; the real complexity comes with scaling it. On 1 to 1000, hosts Jack Altman and Erik Torenberg dig deep into the inevitable twists and turns operators encounter along the journey of turning an idea into a business. Hear all about the tactical challenges of scaling from the people that built up the world’s leading companies like Stripe, Ramp, and Lattice. Our first episode with Eric Glyman of Ramp is out now: https://link.chtbl.com/1to1000
RECOMMENDED PODCAST:
Run the Numbers is a weekly podcast about financial metrics and business models, designed for ambitious people operating tech startups. It's a collection of things host CJ Gustafson (CFO at Partstech and writer of Mostly Metrics) has learned and thought about in the trenches as a tech CFO. Subscribe to listen on the platform of your choice: https://link.chtbl.com/runthenumbers
TIMESTAMPS:
(00:00) Episode Preview
(00:00:56) Introducing Vivek Natarajan and Tao Tu
(00:04:18) The story of Google’s Medical AI research progress
(00:07:11) Multi-modal Med-PaLM
(00:10:32) Genomic data - how do you represent it?
(00:11:13) Google’s Deep Variant
(00:14:44) The successes and failures behind the incredible pace of progress
(00:15:02) Sponsors: Netsuite | Omneky
(00:21:54) Google’s research culture and assembling an interdisciplinary team
(00:31:36) Google’s Pathways
(00:33:40) Med-PaLM M's architecture
(00:37:28) Working with 3 different model sizes and what you learn
(00:46:56) Data and compute required for Med-PaLM M
(00:49:38) Med-PaLM M's cycle time
(00:54:56) Is a bridge or adapter structure worth implementing?
(01:00:09) Can we create an AI doctor?
(01:02:39) Emergent capabilities like identifying tuberculosis
(01:09:37) Reactions to these emergent capabilities
(01:11:13) Moving towards clinical trials and real-world testing
(01:13:01) Regulatory and safety considerations
(01:15:03) AI safety in the healthcare domain
(01:17:00) Potential to transform healthcare access worldwide
LINKS:
Med-PaLM: https://sites.research.google/med-palm/
Med-PaLM M paper: https://arxiv.org/abs/2307.14334
Our earlier conversation with Vivek Natarajan on Med-PaLM: https://www.youtube.com/watch?v=nPBd7i5tnEE
X/TWITTER:
@vivnat (Vivek)
@taotu831 (Tao)
@labenz (Nathan)
@eriktorenberg
@CogRev_Podcast
SPONSORS: NetSuite | Omneky
NetSuite has 25 years of providing financial software for all your business needs. More than 36,000 businesses have already upgraded to NetSuite by Oracle, gaining visibility and control over their financials, inventory, HR, eCommerce, and more. If you're looking for an ERP platform ✅ head to NetSuite: http://netsuite.com/cognitive and download your own customized KPI checklist.
Omneky is an omnichannel creative generation platform that lets you launch hundreds of thousands of ad iterations that actually work customized across all platforms, with a click of a button. Omneky combines generative AI and real-time advertising data. Mention "Cog Rev" for 10% off.
