AI Doctors Beat Real Ones 4-to-1

+ Try On Clothes with AI 🕶️

Microsoft's AI achieves 85% diagnostic accuracy vs 20% for doctors, Baidu declares open-source war on OpenAI, and Google's new virtual styling app...

In this edition we’ll be covering…

  • Microsoft's medical AI that outperforms doctors 4-to-1

  • Baidu's strategic open-source attack on the AI industry's pricing model

  • A tutorial on Google's new AI fashion app Doppl for virtual outfit try-ons

  • 5 trending AI signals including MrBeast's AI controversy and Salesforce's AI workforce

  • 3 trending AI tools to boost your productivity

  • And much more…

The Latest in AI

Microsoft’s AI Beats 85% of Doctors

Microsoft just dropped some mind-blowing medical AI research that should make every doctor take notice. Their new AI Diagnostic Orchestrator (MAI-DxO) correctly diagnosed 85% of complex medical cases from the New England Journal of Medicine, a performance rate that's more than four times higher than experienced physicians who only managed 20% accuracy.

The research team put their AI through the ultimate medical school stress test, using 304 of the most challenging diagnostic cases published in NEJM. And yeah, we're talking about medical mysteries that leave even specialists scratching their heads for years…

What makes this impressive:

  • Instead of just guessing, the AI systematically asks follow-up questions and orders tests, just like a real doctor would during a step-by-step diagnostic process.

  • Not only did MAI-DxO get more diagnoses right, it did so while spending less money on unnecessary tests compared to human physicians.

  • Rather than relying on a single AI model, the system coordinates multiple AI models (including OpenAI's o3) to create a "virtual panel of physicians" with diverse diagnostic approaches.

So What?

Microsoft is creating a system that could fundamentally reshape how healthcare works.

With US health spending approaching 20% of GDP and up to 25% of that being wasted on ineffective treatments, an AI that can diagnose better while costing less could be transformational.

We're looking at a future where AI doesn't replace doctors but supercharges them with near-supernatural diagnostic abilities. So, the real question is, how quickly can we get it into hospitals..?

Get Your Hands Dirty!

Style Yourself with Google’s New AI Fashion App

Google just launched Doppl, an experimental AI app that's like having a personal stylist in your pocket.

The app uses AI to show you exactly how different outfits would look on your body, and it can even animate you wearing them.

Here's how to get started with Doppl:

  1. Download Doppl from the App Store or Google Play (US only for now).

  2. Upload a full-body photo of yourself, make sure you're alone in the frame and the lighting is good.

  3. Take photos or screenshots of any outfit you see online, in stores, or on friends. The app works also with images from Pinterest, shopping sites, social media, literally anywhere.

  4. Upload these outfit photos to the app.

  5. Select an outfit photo and let Doppl's AI map it onto your body. The app generates a realistic image of you wearing the outfit.

  6. Hit the Animate button to create a short video showing how the clothes move on you!

🔥 Pro Tip: The app works best with high-quality, well-lit photos where you can see your full body clearly. Natural lighting and form-fitting clothes in your base photo will give you the most accurate results.

Industry Intel

China’s AI Arms Race Gets Real…

Baidu just hit the nuclear button in the AI wars by open-sourcing their flagship Ernie chatbot. The Chinese search giant is essentially giving away their crown jewels for free, turning the global AI landscape into a battlefield where performance becomes a commodity and pricing becomes the weapon.

This friends, is a calculated attack on the entire business model that's made OpenAI and Anthropic billions. Baidu has been aggressively slashing prices all year, dropping API costs by 80% and making their models free for consumers.

The strategic implications are massive:

  • By offering comparable performance for free, Baidu is forcing Western AI companies to justify their expensive subscription fees and API costs.

  • Open-sourcing means Baidu can tap into AI expertise worldwide, potentially sidestepping US export restrictions and sanctions that limit their access to cutting-edge hardware.

  • This comes as competitors like DeepSeek face delays due to GPU export restrictions, giving Baidu a window to capture developer mindshare globally.

So What?

Baidu is essentially declaring economic warfare on the AI industry's premium pricing model. When a company with 23 million monthly active users decides to give away what others charge premium rates for, it forces everyone to rethink their strategy.

While costs to use AI models are dropping overall, this essentially becomes a race to the bottom. OpenAI might have a solution to compete with this (if they ever drop their open source model).

Baidu's open-source move has the potential to create an even bigger disruption than DeepSeek, forcing Western AI companies to justify why anyone should pay for what's now available for free.

Quick Bites

Stay updated with our favorite highlights, dive in for a full flavor of the coverage!

Meta continues its OpenAI talent raid by hiring four more researchers: Shengjia Zhao, Jiahui Yu, Shuchao Bi, and Hongyu Ren, adding to the growing exodus from Sam Altman's company as competition for AI talent intensifies.

MrBeast pulled his controversial AI thumbnail generator after major backlash from creators who accused him of building a "plagiarism machine" that stole their styles without permission. The $80/month tool was replaced with links to human artists for commissions.

ChatGPT solved a 5-year medical mystery in seconds when a patient with chronic jaw pain tried the AI after countless doctor visits failed. The chatbot suggested a simple exercise that immediately fixed the problem.

Khan Academy CEO Sal Khan predicts AI classrooms will be like having "4-5 amazing graduate students" assisting each teacher, helping with everything from grading papers to providing real-time insights about student engagement and personalized lesson planning.

Salesforce CEO Marc Benioff revealed that AI now handles 30-50% of the company's work with 93% accuracy, calling it a "digital labor revolution" that's allowing employees to focus on higher-value tasks while dramatically boosting productivity.

🚀 Wix Business Launcher Custom GPT - A specialized Custom GPT that helps entrepreneurs generate and validate business ideas with AI-powered market research and planning assistance.

📞 Jamy.ai - An AI meeting assistant that automatically records, transcribes, and summarizes your calls in 50+ languages, with smart task assignment and integration with your favorite productivity tools.

Dify - An open-source platform for building AI applications and workflows, giving developers powerful tools to create custom AI solutions.

The Neural Network

Our last poll revealed that many of you are getting into AI workflow automation!

Almost 2/3s of you guys voted "Summarize today's Slack messages and create action items" when asked how you are going to use 11ai. Honestly, we’re going to be doing very similar things…

On the theme of AI and healthcare, we still think AI is too agreeable to be a standalone, functioning doctor.

The following image helps put things into context…

Until we Type Again…

Thank you for reading yet another edition of Digestible AI. Be sure to give us a follow on X, Instagram, and LinkedIn too!

How did we do?

This helps us create better newsletters!

Login or Subscribe to participate in polls.

If you have any suggestions or specific feedback, simply reply to this email. Additionally, if you found this insightful, don't hesitate to engage with us on our socials and forward this over to your friends!