📉 The End of OpenAI?

On OpenAI Dev Day, Elon's PromptIDE, Meta/HuggingFace, Adept's new tool, and Motion Brush by Runway

Author

Lior Sinclair
November 20, 2023

AlphaSignal

Hey ,

Welcome to this week's edition of AlphaSignal.

Whether you are a researcher, engineer, developer, or data scientist, our summaries ensure you're always up-to-date with the latest breakthroughs in AI.

Let's get into it!

Lior

In Today’s Email:

  • Top Releases and Announcements in the Industry

  • OpenAI board fires Sam Altman, gets hired by Microsoft

  • NVIDIA Introduces Nemotron-3 8B

  • Scientists News Focus on Trillion Parameters Models

Read Time: 4 min 47 sec

RELEASES & ANNOUNCEMENTS

Meta’s Emu Video is a new AI tool for creating 4-second, 512x512 resolution videos at 16 FPS from text prompts. It uses a two-step process: first generating an image from text, then adding motion, enhancing video generation efficiency. Available now.

Multimodal LLMs, for everyone, for free. That is the promise of Kyutai, a new Paris-based non-profit AI lab co-founded by former Google DeepMind talent. It aims to develop muodal large language models (LLMs) accessible to everyone for free, prioritizing scientific advancement without distractions.

The model excels at generating high-quality music with instrumentals and vocals, performing transformation and continun tasks, and giving users more nuanced control of the output’s style and performance. In partnership with YouTube, DeepMind will run two new AI music experiments: Dream Track and Music AI tools.

With 141 GB memory, NVIDIA's H200 Tensor Core GPU achieves nearly double the inference speed for generative AI inference applications compared to the H100. The AI accelerator is set for release in Q2 2024.

At their Ignite conference, Microsoft introduced new models to the Azure AI model catalog, including Nemotron-3 8B, Code Llama, and Mistral. They also introduced 'Models as a Service' (MaaS) for easier AI model integration and customization.

Hire a world-class AI team for 80% less

Building an AI product is hard. Engineers who understand AI are expensive and hard to find. And there's no way of telling who's legit and who's not.

That's why companies around the world trust AE Studio. We help you craft and implement the optimal AI solution for your business with our team of world class AI experts from Harvard, Stanford and Princeton.

Our development, design, and data science teams work closely with founders and executives to create custom software and AI solutions that get the job done for a fraction of the cost.

Book a free consultation session now.

NEWS
OpenAI Board Fires Sam Altman, Replaces with Twitch Founder, Emmett Shear

What's New?
OpenAI has seen dramatic leadership changes: CEO Sam Altman was fired, co-founder Greg Brockman resigned, and CTO Mira Murati became interim CEO. This was followed by the appointment of Twitch co-founder Emmett Shear as another interim CEO.

Subsequently, Microsoft hired Altman and Brockman to head a new AI research unit. Furthermore, three senior OpenAI executives resigned, and over 500 employees signed a letter demanding the board's resignation.

Why Does It Matter?
These shifts are significant for the AI research community as they could influence the direction of future AI developments. OpenAI, known for its groundbreaking work in generative AI systems like ChatGPT and DALL-E, might see a shift in its research and commercialization approach under Shear.

Key Takeaways:

  • OpenAI CEO Altman fired, Brockman resigns; Microsoft hires both for new AI research unit.

  • Emmett Shear’s appointment as OpenAI’s interim CEO may signal a shift in research and commercial strategies.

  • Over 500 OpenAI employees voiced concerns, reflecting potential internal challenges and discontent.

Still struggling to bill for AI and LLM tools? Leave it to Orb.

Charging by tokens or credits? Whichever pricing model you choose - package, matrix, tiered - Orb makes it easy to implement.

Companies like Vercel, Replit, and Airbyte trust Orb to track consumption, prevent fraud, and align pricing to value (even down to GPU runtime).

Special Offer:
The first 10 qualified AlphaSignal readers to sign up get a free trial for Orb.

NEWS
NVIDIA Introduces Nemotron-3 8B to Revolutionize Enterprise AI Development

What's New?
NVIDIA has introduced the Nemotron-3 8B family, a set of generative foundation models within its NeMo framework. This family of models is designed to improve enterprise AI applications, and includes base, chat, and question-and-answer checkpoints.

The Nemotron-3 8B models are available on the Azure AI Model Catalog, HuggingFace, and the NVIDIA AI Foundation Model hub, and can be fine-tuned for custom use cases.

Why Does It Matter?
Nvidia’s NeMo framework enables enterprises to quickly implement AI applications across various environments. The Nemotron-3 8B family gives developers an easy way to integrate foundation models and fine-tuned versions of these models into enterprise-specific requirements.

Main Takeaways

  • Versatile AI Models: Nemotron-3 8B supports extensive languages and coding options, suitable for diverse enterprise needs.

  • Customizable Solutions: It provides base, Q&A, and 3 chat model variants — SFT, RLHF, and SteerLM — allowing for diverse user interactions.

  • Optimal Performance: Integration with TensorRT-LLM and NVIDIA Triton Inference Server ensures low-latency, high accuracy inference.

  • Comprehensive Framework: NVIDIA's NeMo allows easier model building, customization, and deployment.

NEWS
Scientists Join Forces to Focus on Trillion Parameters Models

What's New?
The Trillion Parameter Consortium (TPC) is a collaborative effort aiming to develop AI models with over a trillion parameters. This initiative stands out for its focus on large-scale model architecture and data management, utilizing exascale computing resources. It brings together expertise from various sectors, including federal labs, academia, and industry, to tackle the practical challenges of scalable AI development.

Why Does It Matter?
For those in the field, TPC represents a practical approach to surmount existing limitations in AI model training and data processing. Its emphasis on optimizing AI libraries for exascale computing and developing effective evaluation methodologies addresses some of the key technical challenges in advancing AI applications in scientific research.

Main Takeaways

  • Global Collaboration: TPC builds a community of researchers to tackle the complexities of building large-scale AI models.

  • Focus on Trustworthy AI: Emphasizes developing scalable, reliable AI models for scientific tasks.

  • Interdisciplinary Approach: Aims to create models that integrate knowledge across various scientific fields.

  • Leveraging Exascale Computing: Utilizes emerging exascale computing platforms for training large language models.

Read Last Week’s Summaries:

How was today’s email?

Not Great      Good      Amazing

Thank You

Igor Tica is a contributing writer at AlphaSignal and a research engineer at SmartCat, focusing on computer vision. He's actively seeking partnerships in self-supervised and contrastive learning.

Jacob Marks is an editor at AlphaSignal and ML engineer at Voxel51, is recognized as a leading AI voice on Medium and LinkedIn. Formerly at Google X and Samsung, he holds a Ph.D. in Theoretical Physics from Stanford.

Want to promote your company, product, job, or event to 150,000+ AI researchers and engineers? You can reach out here.