An entirely new way to present ideas

Gamma’s AI creates beautiful presentations, websites, and more. No design or coding skills required. Try it free today.

Today’s top AI Highlights:

1. OpenAI’s co-founder Schulman leaves for Anthropic

2. Run Llama 3.1 on Samsung Galaxy and iPhone with PyTorch’s library

3. Character AI’s low-code platform for prompt designing

4. Elon Musk is suing OpenAI and Sam Altman, again

5. Nvidia scraped ‘A Human Lifetime’ of videos a day to train AI

& so much more!

Read time: 3 mins

Latest Developments

OpenAI Leadership in Turmoil Again ⚠️

OpenAI’s leadership shows no signs of stability since CEO Sam Altman was ousted last year. Several employees from the AI safety team have resigned, followed by co-founder Ilya Sutskever, OpenAI is seeing a row of key departures. Now, John Schulman, a co-founder of OpenAI, has left the company to join rival AI startup Anthropic. At the same time, OpenAI president Greg Brockman announced he will be taking an extended leave until the end of the year.

Key Highlights:

1. John Schulman - While not throwing mud at OpenAI’s efforts towards AI safety, Schulman has joined Anthropic to “deepen his work in AI alignment”. He stated, “I believe I can gain new perspectives and do research alongside people deeply engaged with the topics I’m most interested in.”

2. Greg Brockman - After nine years at OpenAI, Brockman is taking a leave to “relax and recharge,” stepping back from his role temporarily.

3. Another departure - Peter Deng, a product manager who joined OpenAI last year after leading products at Meta, Uber, and Airtable, also exited some time ago, the company confirmed.

4. Why this matters - This change in leadership follows the resignation of the AI alignment team at OpenAI and Sutskever’s new venture SSI. Schulman’s exit to focus on AI safety at Anthropic also suggests that OpenAI’s strategic priorities have shifted. Further, we might see Anthropic’s technology safer and stronger than OpenAI’s as key talents of OpenAI keep shifting there.

PyTorch’s Library to Run LLMs on Laptop, Desktop and Mobile 📱

PyTorch has just launched torchchat to run LLMs like Llama 3.1 smoother and faster locally on devices, including laptops, desktops, and even mobile phones. Torchchat builds on previous PyTorch work and offers key features like export, quantization, and evaluation tools, making it easier for developers to build local LLM inference solutions. The project offers flexibility with Python, C++, and mobile device compatibility.

Key Highlights:

1. Performance Across Platforms - Torchchat demonstrates impressive performance results with Llama 3 8B across hardware configurations and quantization levels. It achieves > 8T/s on the Samsung Galaxy S23 and iPhone.

2. Multiple Deployment Options - The library supports Python through a REST API accessible via CLI or browser, desktop deployment, and mobile deployment using ExecuTorch for on-device inference.

3. Open Source - Torchchat allows you to experiment with the code, provide feedback, and contribute by adding new models, hardware support, quantization schemes, and performance enhancements.

Prompt Poet to Craft Smarter Prompts, Faster 📝

Character.AI built Prompt Poet to make it easier to write complex prompts for large language models (LLMs). Instead of messing around with basic f-strings, Prompt Poet uses YAML and Jinja2, making it more powerful and organized. This helps them handle the huge number of prompts they create every day. It also makes it easier for both developers and non-developers to work on and improve prompts.

Key Highlights:

1. Write Prompts Like Code - Prompt Poet uses templates with YAML and Jinja2, so you can write prompts like you would code, making them more adaptable and reusable.

2. Faster Responses - Prompt Poet helps make sure your prompts run quickly by optimizing how they're processed and cached.

3. Build Modular Prompts - You can break down your prompts into reusable chunks, making them easier to manage and test different versions.

Quick Bites

1. AI startup Groq Inc. has raised $640 million in a new funding round, soaring its valuation to $2.8 billion. The round was led by BlackRock Inc. funds and investment arms of Cisco and Samsung Electronics. Interestingly, Yann LeCun, Meta’s chief AI scientist, will become an adviser to the startup.

2. NVIDIA’s AI team reportedly scraped troves of copyrighted content for training its AI models. The company asked workers to download videos from YouTube, Netflix and other platforms to develop commercial AI projects. NVIDIA has defended this practice by claiming compliance with copyright laws.

3. Elon Musk is again suing OpenAI, Sam Altman, and Greg Brockman, alleging they manipulated him into co-founding OpenAI under false pretenses and breached the company’s founding mission. The new lawsuit also claims OpenAI’s deal with Microsoft breaks federal racketeering laws and defrauds Musk.

4. Meta is accepting applications for the Llama 3.1 Impact Grants, offering up to $500,000 for projects using Llama 3.1 to tackle social challenges. Submit proposals by November 22, 2024, and participate in regional events for additional funding.

Tools of the Trade

1. Pinokio: A browser that lets you locally install, run, and automate any AI models on your computer. Everything you can run in your command line can be automated with Pinokio script, with a user-friendly UI.

2. Hero: All-in-one AI daily assistant app that combines calendars, reminders, notes, weather updates, and grocery orders. You can give voice inputs to it for quick updates.

3. Udio v1.5: AI music generation app with enhanced audio quality, key control, and improved global language support, along with new features like a dedicated creation page, stem downloads, audio-to-audio remixing, and shareable lyric videos.

4. Awesome LLM Apps: Build awesome LLM apps using RAG for interacting with data sources like GitHub, Gmail, PDFs, and YouTube videos through simple texts. These apps will let you retrieve information, engage in chat, and extract insights directly from content on these platforms.

Hot Takes

1. I went to a techie party in the hart of the Silicon Valley this weekend. Talked to several Google employees among others. Not a single one of them had heard of Kaggle. ~
   Bojan Tunguz
   

2. Good programmers understand that simple code is always better than clever code.
   Writing clever code doesn't make you smart. It makes you a fool. ~
   Santiago

Meme of the Day

That’s all for today! See you tomorrow with more such AI-filled content.

Real-time AI Updates 🚨

⚡️ Follow me on Twitter @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!

PS: We curate this AI newsletter every day for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!