Howdy! Welcome! This is the “AI Daily” newsletter—your daily dose of AI news and resources.

In today’s edition:

• 🔌 OpenAI’s AI chip update

• 🤹 GitHub Copilot goes multi-model

• 🤖 Plus trending AI tools, posts, and resources.

Ready, set, go…

🔌 OpenAI’s AI chip update

OpenAI is reportedly working with Broadcom and TSMC to develop its first in-house chip for handling large AI workloads, with production possibly starting in 2026.

Here's what you need to know:

• It’s been months since work began on a custom chip, with a team of about 20 people, including former Google engineers who developed Tensor processors.

• Microsoft Azure, OpenAI's key infrastructure partner, will deploy AMD chips in OpenAI's systems, helping diversify its hardware portfolio beyond Nvidia processors.

• OpenAI initially considered building its own network of chip manufacturing facilities, known as "foundries," but paused the plan due to high costs and time demands.

OpenAI’s chip strategy unveiled:

• Hybrid Approach: OpenAI combines internal efforts with external partnerships to address chip shortages and manage costs, mirroring strategies of Amazon, Meta, Google, and Microsoft.

• Partnerships: Working with Broadcom and TSMC to expand chip supply options in a market dominated by Nvidia.

• Nvidia Relations: OpenAI remains cautious to avoid straining its relationship with Nvidia, especially as it seeks access to their next-gen Blackwell chips.

With the AI chip industry expected to grow to nearly $165 billion in revenues by 2030, the push for proprietary technology has become more competitive than ever.

🤹 GitHub Copilot goes multi-model

GitHub is expanding its Copilot Chat to support multiple code-completion models and is also launching Spark, a new AI tool to help anyone build web apps.

Here's what you need to know:

• GitHub Copilot users can now switch between the multiple latest AI models, including Anthropic’s Claude Sonnet 3.5, Google Gemini 1.5 Pro, and OpenAI’s GPT-4, during a session to better meet their coding needs.

• Spark, GitHub's new AI tool, allows users to build simple web apps with plain English prompts, offering instant live previews and enabling interactive adjustments through conversations with the bot.

• Once the app is created, users can run it on a desktop, tablet, or mobile device and also share the app with others to let people remix and build on top of Spark apps.

• Other notable new features include the ability for Copilot in VS Code to edit multiple files at once and a code review feature that gives feedback in 30 seconds.

Quick Hits

1. Google CEO Sundar Pichai reveals that over 25% of the company's new code is created by AI, with human engineers reviewing and approving the machine-created code (1 minute read)

2. OpenAI introduced a new search function for the ChatGPT web app that helps you search through your chat history, making it easy to find previous conversations or pick up where you left off (X Post)

3. Salesforce announced the general availability of Agentforce, a platform that lets companies create customizable AI agents to automate tasks across different business functions (5 minutes read)

4. OpenAI and Anthropic revenue breakdown. OpenAI's revenue is at an annualized run rate of ~$3.6B as of August 2024, up from $1.6B in 2023. Anthropic's revenue is projected to reach $1B by year-end, a 900% increase from ~$100M (5 minutes read)

5. Market share report of top Generative AI chatbots (ChatGPT, Google Gemini, Perplexity, and ClaudeAI) by October 2024. ChatGPT remains the market leader, but its growth has eased as both Google and Microsoft release improvements (4 minutes read)

6. Elon Musk urges users to submit X-ray, PET , MRI or medical images to xAI chatbot Grok for analysis. This is still early stage, but it is already quite accurate and will become extremely good (X Post)

Useful AI Links

Trending Tools

• Midjourney’s SREF Codes > Explore 498 unique MidJourney Sref codes and style reference prompts. (link)

• Create > A new text-to-app AI code builder. Seems very promising. (link)

• Wikiwand > AI-driven contextual learning with Wikipedia. (link)

• MindPalace > Integrate your data from other apps and create your own personal time machine. (link)

Resources / Guides

• Building enterprise AI apps with multi-agent RAG systems (MARS). (link)

• I own my LLM chat history, and so should you. (link)

• Meta’s AI abundance. (link)

• The AI productivity paradox - Why aren’t more workers using ChatGPT? (link)

Daily Dose Of Contents

New ’Robot Dreams’ movie that will double the size of your heart. (1 minute trailer)

Elon Musk and Dr. Peter Diamandis FII8 conversation on the future of AI. (22 minutes video)

That’s All For Today, Folks

Thanks again for being here.