🚑 AI in Healthcare with AMIE - 🤖 This Week in AI - ⚖️ Balancing Talents

Welcome back to 3-in-3, your turbo-charged, knowledge-packed espresso shot for the busy, tech-savvy, future-gazing brainiacs who crave to outsmart tomorrow, today, and all in the time it takes to microwave popcorn!

Today on the menu:
Thing 1 - Google's AMIE and Its Impact on Healthcare
Thing 2 - AI of the week
Thing 3 - Navigating Multiple Talents

Thing 1 - Google's AMIE and Its Impact on Healthcare

AMIE is an AI system based on large language models (LLMs) that can perform diagnostic reasoning and dialogue with simulated patients.

AMIE beat human doctors based on "patients"' rated quality of care.

Screenshot of AMIE conversation

Ironic, isn't it? AI that doesn't have empathy does a better job showing it than humans who have it...

AMIE accuracy graph from Google Research

In addition to strong standalone performance, using the AMIE system led to significant assistive effect and improvements in diagnostic accuracy of the clinicians in solving these complex case challenges.

AMIE helped doctors make better decisions on diagnostics, but... it did better on its own 🫠 😬

Might be too good to be true (for now)

• The study was not fair to the doctors, because they had to use text chat instead of talking.
• The research is a first step, and there is a lot more work to do before the AI system can be used in real life.
• There are many things to think about, like making sure the AI system is fair, safe, private, and does not make mistakes.

Thing 2 - AI of the week

I've seen more research-based AI stuff this week than practical and available. Which means more will come soon too 😉

• Luma AI dropped a text-to-3D generator on iOS and Web
• Littercam uses AI to spot littering from vehicles (not new but new to me)
• Rabbit Tech announced (and sold out) R1 - their pocket companion

Screenshot of R1 from Rabbit website

• OpenAI launched GPT Store to ChatGPT+ subscribers
• FairOffer - estimates fair compensation — for any title, at any stage of the company, anywhere in the world
• Google's new medical LLM - AIME - beat human doctors based on patients' rated quality of care

Thing 3 - Navigating Multiple Talents

There are many smart and talented people who struggle to execute their long-term vision precisely because they can do so many different things: they get distracted by everything, do nothing. People who get things done simply know how to focus, how to commit, how to make a choice.

- Orange Book

How do multiple things you do fit into your long-term vision?
Are there things you do that don't and you can drop them?

More focus going forward, Z 🙂

Cheers, Zvonimir