- AI Valley
- Posts
- ChatGPT Outperforms Human Doctors
ChatGPT Outperforms Human Doctors
PLUS: Perplexity AI online shopping
Together with
Howdy! It’s Barsee again.
Happy Tuesday, AI family, and welcome back to AI Valley.
In today’s edition:
ChatGPT Outperforms Human Doctors in Medical Diagnosis
Mistral launches Pixtral large and upgrade its Le Chat Assistant
Plus trending AI tools, posts, and resources
Ready, set, go…
CHATGPT
ChatGPT outperforms human doctors in medical diagnosis
A recent study showed that ChatGPT outperformed human physicians in assessing medical case histories, even when the doctors used a chatbot.
Here's what you need to know:
Dr. Adam Rodman tested 50 physicians to see if ChatGPT improved diagnostic accuracy.
ChatGPT-4 achieved 90% accuracy in diagnosing and explaining conditions.
Doctors using ChatGPT-4 scored 76%, slightly better than the 74% of those not using it.
Physicians often stuck to their original diagnoses, ignoring better suggestions from ChatGPT.
The study highlighted a gap in doctors’ use of AI tools for complex diagnostics.
Key Takeaway:
AI could improve diagnostic accuracy and patient outcomes, but more research is needed to refine human-AI collaboration.
TOGETHER WITH GROWTHSCHOOL
2 hour LinkedIn + AI Workshop: The Secret Weapon for Founders, C-Suites & Top-Level Professionals
LinkedIn isn’t just a social platform, it’s a goldmine when you combine it with AI.
Vaibhav Sisinty, with over 400K followers on LinkedIn and 1 million followers across social media, has taught 200,000+ professionals how to unlock this potential.
In his AI Powered LinkedIn Workshop, Vaibhav will show you how to:
Automate lead generation to grow your business while you sleep
Leverage AI to land high-paying jobs without wasting hours on applications
Master his $100K LinkedIn Outbound Strategy to boost revenue effortlessly
Use AI to create and distribute content, saving you hours every week
This workshop is the real deal for anyone who wants to dominate LinkedIn in 2024 and beyond.
Claim Your Free Spot Now (Only 100 free seats!)
SIDE UPDATES
Perplexity has launched new shopping features, including “Buy with Pro,” allowing users to purchase directly through the platform with free shipping on eligible orders. The "Snap to Shop" feature lets you upload a photo to find product options, and all recommendations are unsponsored for more genuine suggestions. Currently available in the U.S., Perplexity plans to expand. Just remember to double-check prices, as AI can still make mistakes.
Coca-Cola tasked three creative agencies with reimagining their legendary "Holidays Are Coming" advert using artificial intelligence. The brief included recreating the iconic illuminated truck and exploring hyper-realistic visuals and innovative storytelling. The initiative highlights how AI is reshaping the future of creative production.
A cutting-edge AI learning method, test-time training, has achieved a significant breakthrough in AGI benchmarks. It scored 61.9% on the ARC Prize, surpassing the average human performance of 60.2% and establishing a new state-of-the-art in the field. This milestone highlights the growing capabilities of AI systems in tackling abstract reasoning tasks traditionally dominated by humans.
Nvidia has introduced LLaMA-Mesh, a groundbreaking extension of large language models (LLMs) capable of generating and understanding 3D mesh objects. This innovation pushes LLMs beyond text into the realm of 3D modeling. The system uses the OBJ format, a widely recognized text-based standard for representing 3D objects, marking a significant step in bridging AI and 3D design.
Spotify has introduced a new feature allowing users to create playlists by simply typing prompts. The tool generates a 30-song playlist based on the input and lets users refine it further with additional prompts, such as requesting more upbeat tracks. Users get better results by specifying genres, decades, moods, or favorite artists, making it easier to match playlists to a desired vibe.
ElevenLabs has introduced tools to create conversational AI bots with customizable tone, response length, and LLM integration, including GPT and Claude. Users can add knowledge bases, adjust latency, and integrate custom models using its SDK. The company is also developing speech-to-text capabilities, positioning itself against OpenAI and Google.
MISTRAL
Mistral launches Pixtral large and upgrade its Le Chat Assistant
Mistral has unveiled Pixtral Large, a visual model that sets new benchmarks, and upgraded its Le Chat assistant with web search and image generation features.
Here's what you need to know:
Pixtral Large, based on Mistral Large 2, has a 128K context window and can process up to 30 high-res images at once.
It scored 69.4% on MathVista, surpassing GPT-4 and Gemini 1.5 Pro in visual math reasoning.
The model also outperformed Claude 3.5 Sonnet and others in diagram analysis and real-world tasks.
Le Chat now includes web search and can generate images via Flux Pro mode from Black Forest Labs.
With Pixtral Large, Le Chat can analyze complex PDFs, handling tables, diagrams, and formulas.
Why it matters:
These updates position Le Chat as a powerful assistant for both visual and text-based tasks, enhancing AI's multimodal capabilities.
TRENDING TOOLS
MagicQuill > A new lightweight AI Image Editor that’s way faster than Photoshop. [Link]
Locofy - AI-powered frontend development with lightning-fast design-to-code conversion. [Link]
GenSpark Finance > Uses Claude AI to turn complex financial data into easy-to-understand visual reports. [Link]
Gecko Security > AI-powered security engineer that finds and fixes vulnerabilities in your code. [Link]
Napkin > Transforms your existing text content into visuals like diagrams, charts, scenes, and image. [Link]
COOL FINDINGS / RESOURCES
How to run AI models on your phone without any internet connection - by Paul [Link]
I was just informed from my boss and HR that my entire profession is being automated away [Link]
Notion engineering blog [Link]
How you can use HeyGen’s new Avatar 3.0 to create an AI clone of yourself in just 7 steps - by Curious Refuge [Link]
Build website/app backend using AI: Learn how to store data and authenticate users in your bolt.new app - by StackBlitz [Link]
How to create Flowcharts and Diagrams with ChatGPT - by AI for Education [Link]
This paper explores Claude 3.5 'computer use' capabilities across different domains and software [Link]
We studied 200,000 AI Overviews: Here's what we learned - by Semrush [Link]
CONTENT CORNER
The AI models making huge money on Instagram.
A Video-to-Sound Effect AI tool would be a game-changer for AI filmmaking.
A Video-to-Sound Effect AI tool would be a game-changer for AI filmmaking.
This Visually-Indicated-Sounds technology developed by @MIT_CSAIL lets us input a video and automatically generate natural sound based on the visuals.
Sound is just as crucial as visuals in… x.com/i/web/status/1…
— el.cine (@EHuanglu)
6:51 PM • Nov 14, 2024
How AI was stolen.
THAT’S ALL FOR TODAY
That’s all for today’s issue, folks.
💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee
👍️ Like what you see? Subscribe here
Thanks for being here.
HOW WAS TODAY'S NEWSLETTER |
REACH 100K+ READERS
Acquire new customers and drive revenue by partnering with us
Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.
If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.