The AI Files

#25 - AI with Nobel prizes, Tesla's robots, VoiceRAG, AI adoption etc

In this episode, we delve into the Nobel prizes for contribution via AI, Tesla's 'We, Robot' event with launches of Olympus, Cybercab and Robovan as well as new paper on Adoption of AI, VoiceRAG, Azure demo on advanced voice capability and a few other items.
 
Show notes:
AMD sees next AI chip in mass production later this year | Reuters
Cancer AI Alliance joins medical and tech expertise together with $40M to collaborate on next-gen care | TechCrunch
DeepMind's Demis Hassabis and John Jumper scoop Nobel Prize in Chemistry for AlphaFold | TechCrunch
Geoff Hinton and John Hopfield win Nobel Prize in Physics for their work in foundational AI | TechCrunch
Liquid AI, a new MIT spinoff, wants to build an entirely new type of AI
Announcing Azure OpenAI Global Batch Offering: Efficient processing at scale with 50% less cost - Microsoft Community Hub
Unlocking next-generation AI capabilities with healthcare AI models - Microsoft Industry Blogs
Introducing canvas, a new way to write and code with ChatGPT. | OpenAI
Announcing new products and features for Azure OpenAI Service including GPT-4o-Realtime-Preview with audio and speech capabilities | Microsoft Azure Blog
GitHub - Azure-Samples/aisearch-openai-rag-audio: A simple example implementation of the VoiceRAG pattern to power interactive voice generative AI experiences using RAG with Azure AI Search and Azure OpenAI's gpt-4o-realtime-preview model.
VoiceRAG: An App Pattern for RAG + Voice Using Azure AI Search and the GPT-4o Realtime API for Audio - Microsoft Community Hub
https://youtu.be/vXJka8xZ9Ko (VoiceRAG Demo)
 
 

#24 - Interview with ChatGPT, Canvas, Advanced Voice, Copilot refresh, MovieGen etc

In this episode, I interviewed ChatGPT with Advanced Voice Capabilities. Also, discussed OpenAI DevDay launches including Real Time API, Canvas, Prompt Caching & Vision fine tuning along with Copilot refresh from Microsoft and Meta's Movie Gen. Covered major news, UN's AI recommendations and AI company Cerebras.
Show Notes:
Announcing new products and features for Azure OpenAI Service including GPT-4o-Realtime-Preview with audio and speech capabilities | Microsoft Azure Blog
An AI companion for everyone - The Official Microsoft Blog
Devday Hub | OpenAI
The United Nations Wants to Treat AI With the Same Urgency as Climate Change | WIRED
Cancer AI Alliance joins medical and tech expertise together with $40M to collaborate on next-gen care | TechCrunch
Influence of a Large Language Model on Diagnostic Reasoning: A Randomized Clinical Vignette Study (medrxiv.org)
Opinion | I’m a Doctor. ChatGPT’s Bedside Manner Is Better Than Mine. - The New York Times (nytimes.com)
U.S. Hiring Accelerated in September, Blowing Past Expectations (msn.com)
Cerebras Files Prospectus for I.P.O. - The New York Times (nytimes.com)
How Meta Movie Gen could usher in a new AI-enabled era for content creators
California governor vetoes AI regulation opposed by tech industry - The Washington Post
Artificial Intelligence Advisory Board’s Report ‘a Crucial Milestone in Efforts to Ensure AI Serves all of Humanity’, Says Secretary-General, in Video Message | Meetings Coverage and Press Releases (un.org)

#23 - Meta Orion, Age of Intelligence, Klarna, Microsoft's Trustworthy AI, Spotlighting etc

In this episode, we covered Meta's Connect event that had peak preview of Orion AR glasses and comparison to Meta Ray-Ban smart glasses, Meta AI's multi-modal features and integration with Meta's social media platforms, Microsoft's Trustworthy AI, a set of prompt engineering techniques Spotlighting from Microsoft Research along with news and happenings.

Episode 22 - AI's energy needs, regulations, Large World Models, Copilot Wave 2, US Open etc.

In this episode, we discussed AI's energy needs, California regulations, Copilot Wave 2, AI at US Open etc. We also covered World Labs as AI startup of the week and their goals on Large World Models (LWM).

Episode 21 - A talk with Imran Chaudhri from Progress

"Great AI requires great IA", says Imran Chaudhri, HLS chief architect at Progress. In this episode, we had a talk with Imran about AI and the data dependencies as well as the trends and predictions on AI and other fascinating bits. 

Episode 20 - OpenAI's o1 models, AI's impact on elections, Apple Intelligence, Prompt Injection etc

This episode covers o1-preview and o1-mini models, AI's impact on elections, Apple Intelligence, Prompt Injection & prevention and AI start-up 'Learn Prompting' etc.
 
OpenAI introduces a new series of pathbreaking models, o1-preview and o1-mini, designed to enhance reasoning, problem-solving, and critical thinking processes. These textual only models, with a 128K context window, excel in tasks across physics, chemistry, and biology benchmarks, performing akin to PhD students. Key features include complex code & algorithm generation, advanced problem-solving & brainstorming, and precise document comparison, adept instruction following & workflow management making them ideal for scientific, legal, and coding applications.
Here are the highlights:
- Augments, not replaces existing GPT-4 models. These models are not next version of GPT-4 models. This is a new series of models suited for reasoning, problem solving and thinking
- o1 models prioritize thoughtful responses, producing more tokens and are pricier per token - Particularly suitable for scientific, legal, and coding use cases
- Accessible on Azure OpenAI playground for early evaluation. Compare them against GPT-4 models to pick the best model for specific use cases. Models are available same day in Azure OpenAI playground with early access to try in Azure AI Studio
- Inherently step-by-step thinkers, minimizing the need for chain-of-thought prompting
- High resistance to jailbreaking, enhancing security. They score 84 on a scale of 0-100 in terms of hard to jailbreak compared to 22 for GPT-4o
- o1 models take time to think, generate more tokens on average and are more expensive per token than other GPT models so choose them for appropriate use case. o1-mini will be 80% cheaper than o1-preview
- ChatGPT Plus and Team users will be able to access o1 models in ChatGPT but with weekly limits. In future, OpenAI plan to make ChatGPT triage to select the right model based on the type of requests. Hybrid LLMs is a right approach for internet scale applications
- GitHub Copilot, Devin by Cognition and Harvey's professional services platform are the early adopters and endorsers
• Broadcom CEO: Tech Giants’ Shift Toward Custom AI Chips Will Accelerate — The Information https://www.theinformation.com/briefings/broadcom-ceo-tech-giants-shift-toward-custom-ai-chips-will-accelerate?rc=zzu24f
• Bringing generative AI to video with Adobe Firefly Video Model | Adobe Blog https://blog.adobe.com/en/publish/2024/09/11/bringing-gen-ai-to-video-adobe-firefly-video-model-coming-soon
• Introducing o1: OpenAI's new reasoning model series for developers and enterprises on Azure | Microsoft Azure Blog https://azure.microsoft.com/en-us/blog/introducing-o1-openais-new-reasoning-model-series-for-developers-and-enterprises-on-azure/?msockid=05d6d43ff45563632363c6e8f55762a6
• Apple Intelligence Preview - Apple https://www.apple.com/apple-intelligence/
• The Prompt Report https://arxiv.org/html/2406.06608v1
• C2PA https://c2pa.org/

Episode 19 - 'How AI is transforming Media & Sports': a talk with Scott Bounds from Microsoft

In this episode, we talked with Scott Bounds, a Sr Director of Data & AI at Microsoft. His deep experience in the industry of Media & Sports, comes out as rich insights about the industry embracing AI at a rapid speed. 

Episode 18 - Phi3.5, Image GenAI tools, Auto fine-tuning, AI in Marketing etc

In this episode, we cover Zuckerberg-Elk joint piece in The Economist, AI start up Empower for auto fine-tuning, multiple launches of image generation tools like Imagine 3 and Ideogram, new NeMo pruned model, Phi3.5 launch and Text-To-Speech Avatars from Azure as well as how AI is transforming enterprise functions. 

Episode 17 - Grok 2, AI risks & regulations, Gemini Live, Walmart AI etc

In this episode, we covered a number of news & announcements related to Grok 2, Azure self-serve PTUs, AI Risks & Regulation, Prompt Caching and Walmart using AI etc.
 
Show Notes
MIT researchers release a repository of AI risks | TechCrunch
The AI Risk Repository: A Comprehensive Meta-Review, Database, and Taxonomy of Risks From Artificial Intelligence (website-files.com)
AI Risk Repository (mit.edu)
(20) lmsys.org on X: "Exciting Update from Chatbot Arena! The latest @OpenAI ChatGPT-4o (20240808) API has been tested under "anonymous-chatbot" for the past week with over 11,000 community votes. OpenAI has now successfully re-claimed the #1 position, surpassing Google's Gemini-1.5-Pro-Exp with an https://t.co/gxCDuBOi9N" / X
Models - OpenAI API
Prompt caching with Claude \ Anthropic
California AI Bill Heads to State Assembly — The Information
Elevate your AI deployments more efficiently with new deployment and cost management solutions for Azure OpenAI Service including self-service Provisioned | Microsoft Azure Blog
Grok-2 Beta Release (x.ai)
Azure OpenAI Provisioned August 2024 Update - Azure OpenAI | Microsoft Learn

The AI Files - Episode 16

This week, we're diving deep into the latest power shifts at OpenAI, exploring what these executive changes mean for the future of AI. We also cover Amazon's impressive new Titan Image Generator v2 and the introduction of Figure 02, the humanoid robot pushing the boundaries of AI-driven robotics. Plus, we discuss the growing momentum in AI regulation, with new legislation on deepfakes taking center stage. Tune in to stay ahead of the curve in AI innovation and policy!