Can AI coach 100 million Americans to health?

News, AI leaders, business insights and more

This Week’s Line-up

  • Can AI coach 100 million Americans to health?

  • The Good, Bad, and Ugly about AGI

  • Woman in AI: Meet Suchi Saria

  • Case Study: Waymo and the Robotaxi Revolution

  • Pro Tip: Open vs. Closed Source - A Simple Guide

ps. Connect with me on LinkedIn and Instagram for daily news, insights, and more. 😎

NEWS YOU CAN’T MISS

Can AI coach 100 million Americans to health?

Arianna Huffington and OpenAI CEO Sam Altman have joined forces to create Thrive AI Health, a new venture aimed at developing an AI-powered health coach.

130 million Americans live with at least one chronic condition.

Thrive AI’s health coach will offer them personalized health suggestions based on individual biometric data, lab results, lifestyle choices and calendars.

Recommendations are provided across five key areas: sleep, nutrition, exercise, stress management, and social connections.

DeCarlos Love, a former Google executive with expertise in wearable technology and product development, has been appointed as the CEO of Thrive AI Health.

There are quite a few startups in this space but Huffington's wellness expertise paired with Altman's AI prowess gives Thrive AI Health an unbeatable edge. No other health coach offers this unique blend of deep health network and cutting-edge AI personalization.

I've always admired Arianna Huffington's bold vision first with her media venture Huffington Post, which she sold to AOL and then with her wellness venture Thrive Global. She also wrote the first book I read on the importance of sleep (a topic everyone now talks about).

ps. Altman is 39 years old and Huffington is 73 years old proving that age is nothing but a number!

For anyone out there who thinks it's too late to start your own company, note that Arianna:  

  • Co-founded The Huffington Post in 2005 at age 55

  • Sold HuffPost to AOL for $315 million in 2011

  • Founded Thrive Global in 2016 at age 66

  • Co-founded Thrive AI in 2024 at age 73

It's time we stopped ageism in tech! Very glad to see this partnership! 🥳

TREND ALERT

The Good, Bad, and Ugly about AGI

Bloomberg just reported that OpenAI CEO Sam Altman recently discussed achieving Artificial General Intelligence (AGI) with his employees. AGI is an AI system capable of any intellectual task a human can perform - think beyond emails and images, to complex reasoning, learning, and innovation.

Sounds like science fiction, but Silicon Valley is actively pursuing it. Here’s how OpenAI classifies AI progress.

  • Level 1: Conversational AI (current) AI like ChatGPT engaging in human-like conversation.

  • Level 2: Reasoners (coming soon) AI problem-solving at PhD level.

  • Level 3: Agents (future) AI autonomously performing tasks for days.

  • Level 4: Innovators (future) AI assisting in developing inventions.

  • Level 5: Organizations (future) AI performing work equivalent to entire organizations, the ultimate AGI goal.

It can quickly go from good to bad to ugly.

The potential for great advancements in medicine, scientific discovery, and solving global challenges is clear with more sophisticated AI, but so is the potential for misuse or unintended consequences (think autonomous weapons systems, widespread job displacement, or fake news eroding public trust).

OpenAI faces growing scrutiny over safety concerns. Employees have raised alarms, and co-founder Ilya Sutskever resigned over inadequate safety practices.

Most recently, an anonymous source revealed to The Washington Post that OpenAI allegedly rushed safety tests and prematurely celebrated product launches before ensuring adequate safety measures were in place.

Even if we accept AGI, can we trust OpenAI to develop it responsibly?

We should demand that Sam Altman explain his safety model for each AI level.

Sam Altman isn't alone in predicting the rapid approach of AGI. Nvidia CEO Jensen Huang believes it could arrive within five years, performing a broad spectrum of cognitive tasks at or above human levels. Elon Musk is even bolder, suggesting AI could surpass the brightest humans by 2026.

🤔 These predictions, even if optimistic, highlight the urgency of addressing AGI's potential impact.

WOMEN IN AI

Meet Suchi Saria

IMAGE CREDIT: WILL KIRK / JOHNS HOPKINS UNIVERSITY

Celebrating this week's Woman in AI 🥳: Meet Dr. Suchi Saria, Associate Professor and Research Director, Malone Center for Engineering and Healthcare at Johns Hopkins University.

Born in Darjeeling, India, Saria got both her BA in Computer Science from Mount Holyoke College and her MSc and PhD in AI from Stanford University.

Here is a look at her career trajectory:

  • Held roles at IBM, Goldman Sachs and Microsoft Research

  • Advised at companies including Sanofi, Halcyon Health and PatientPing

  • Co-founded Bayesian Health, a health AI startup focused on developing predictive and preventive healthcare tools, in 2018

At Bayesian Health, Saria led work on TREWS – an early warning system using AI for patients of sepsis - which was able to achieve a reduced mortality of 18.2%. It was named as one of TIME’s Best Inventions of 2023.  

According to Saria, the journey of Bayesian Health began with her nephew tragically passing away from sepsis – this acted as inspiration for her to find the startup.

In 2022, Saria co-founded the Coalition for Health AI (CHAI) to encourage the development of transparent and credible AI systems in healthcare.

CHAI, a coalition of healthcare experts and tech giants like Mayo Clinic, CVS Health, Google, Amazon, and Microsoft, is working with regulators such as the FDA and OSTP to develop agreed-upon standards for evaluating AI technologies in healthcare.

Saria has been recognized as:

  • Modern Healthcare's Top 25 Innovator  

  • MIT Technology Review's 35 Innovators Under 35

  • Young Global Leader by the World Economic Forum  

  • AI’s 10 to Watch by IEEE  

  • World Technology Forum’s Technology Pioneer

Saria also serves on the editorial board of the Journal of Machine Learning Research and on the National Academy of Medicine AI Code of Conduct.

If this the decade of AI in Healthcare (as many predict) then I’m really glad we have responsible AI trailblazers like Saria on our side. 😎 

ENTERPRISE AI CASE STUDY

Waymo and the Robotaxi Revolution

Industry: Automotive

A Waymo autonomous taxi in San Francisco. Image: David Paul Morris/Bloomberg

Sometimes we get so excited by generative AI and AGI that we forget that the most tangible and powerful uses of AI are right before our eyes.

Case in point: Google subsidiary Waymo is now offering robotaxis to everyone in San Francisco after years of testing (no more waitlist to try it out!). It’s a huge milestone for a company that’s been testing it’s driverless cars for years and marks a new era in transportation as we know it.

Waymo boasts a fleet of hundreds of cars, providing tens of thousands of weekly driverless taxi rides throughout San Francisco and it’s become a bit of a tourist attraction in the city.

My 72-year old aunt ordered a Waymo taxi in San Francisco and the driverless car rolled up, she rode in it and absolutely loved it. 😊

But let’s not forget, the key to Waymo’s self-driving system is AI.

Waymo Driver is a complex network of AI algorithms and sensors that work together to perceive and navigate the world around them. Here's how it works:

  1. Perception: Sensors including lidar, radar and cameras, gather data which helps AI identify objects. Example: It detects other vehicles, pedestrians, and cyclists.

  2. Prediction: AI forecasts movements of other road users. Example: It anticipates reactions to changing traffic conditions.

  3. Planning: AI charts the safest, most efficient route. Example: It considers traffic conditions, road rules, and vehicle capabilities.

  4. Control: AI directs the vehicle's actions. Example: It commands the steering, brakes, and accelerator for smooth, safe maneuvering.

Safety is a top priority for robotaxis. Waymo points to impressive data: during the first 7 million rider miles, passengers in Waymo's robotaxis experienced 85% fewer injury-causing crashes compared to traditional vehicles.

The company was named one of Time Magazine’s most influential companies in 2024.

But the competition is on.

  • China has expanded its list of top-tier cities permitting driverless robotaxis without safety supervisors to include Shanghai. And they are popular: Baidu's robotaxi Apollo Go offers less than half the cost of traditional taxis, as reported by Huxiu news.

  • Elon Musk, never one to be left behind, says he will be releasing a Tesla robotaxi this year.

  • Special shoutout to Raquel Urtasun, one of our featured Women in AI, who just raised $200m from Nvidia and Uber for AI-powered commercial trucks.

😎 The future of transportation is happening now, and AI is at the wheel.

PRO TIP

Open vs. Closed Source - A Simple Guide

Don’t let tech lingo throw you off. Here’s a quick intro to two commonly used words.

Imagine AI models as recipes for making smart computer programs. There are two main types: open source and closed source. Let's break them down.

Think of Open Source AI models like a family recipe that's been shared with everyone.

  • The recipe (code) is public - anyone can see it, use it, or even change it

  • It's like cooking with friends - many people can help improve the recipe

  • Often free to use, like a potluck dinner

Closed Source AI models are more like a secret recipe from a famous restaurant.

  • The recipe is kept secret - only the chefs (company) know exactly how it's made

  • You can order the dish (use the AI), but you can't see how it's cooked

  • Usually costs money to use, like ordering from a restaurant

As a general rule, people use Open Source when:

  • They want to understand how things work "under the hood"

  • They need to make changes to fit your specific needs

  • They’re on a tight budget and can handle some technical work

And people use Closed Source when:

  • They want the latest and most advanced features

  • They need reliable performance without worrying about technical details

  • They're okay with paying for a service that "just works"

For generative AI models: ChatGPT is closed source, Mistral is open source and Llama is a bit of a hybrid with some degree of openness while maintaining certain restrictions and controls.

😎 Remember, all types are useful. It just depends on your budget, technical ability and how much you need to customize it. My team and I spend a lot of time with our clients explaining the pros and cons of such options when helping them build AI use cases.

See you next week!

-- The future awaits. Ayesha ♥️

I'd love your feedback...

Please vote below to help me improve the newsletter.

Login or Subscribe to participate in polls.