In this episode of "AI Wanna Talk," host Jacob Norgord dives into the latest advancements in AI, exploring practical applications and significant announcements from major tech companies. This episode covers a range of new AI products, features, and research.

OpenAI's Projects Feature:

• Discussion of OpenAI's "Projects" feature, which allows users to upload files and provide explicit instructions to maintain context throughout a conversation with AI models, addressing the common issue of AI forgetting earlier parts of a conversation. This feature is available for ChatGPT Plus or Pro members, or those with a team account.

Google's Gemini 2.0 Model and AI Agents:

• An overview of Google's new Gemini 2.0 family of models, focusing on the "agentic era" where AI acts proactively on the user's behalf.
• Details on the experimental Gemini 2.0 Flash model, a smaller model designed for low latency and integration into various Google experiences, such as Google Sheets, Docs, and Search.
• Explanation of the "whisk" experiment, allowing users to combine objects from multiple images.
• Information on Google's new state-of-the-art video model, Veo 2, a competitor to OpenAI’s Sora.

XAI's Grok and Mainframe's AI Agents:

• Announcement that XAI's Grok AI assistant is now free for all X users, highlighting its ability to access real-time data from the web and the X platform.
• An introduction to Mainframe, a company developing AI agents that work without human intervention, focusing on their first stage rollout called "Cobbot," which will consist of a suite of AI agents to accelerate teams.

AI and Employment:

• Discussion of how the company CLA is using AI to boost productivity and potentially replace roles, and the potential implications for the job market and quality of output.

Particle News App:

• Highlight of the Particle News app, which uses a TikTok-like algorithm for personalized news feeds and includes features like article narration and AI-powered Q&A.

Social Media and Teen Depression:

• Exploration of data from Jonathan Height's "The Anxious Generation", revealing a correlation between the rise of smartphones and social media with the increase in teen depression.

Links:

• Google Labs
• Veo 2 Waitlist
• Grok
• Particle News App
• Twitter thread on "The Anxious Generation"

(Definitely meant 25th power not 25th degree toward the end of the episode)

In this episode of "AI Wanta Talk," host Jacob Norgord explores recent AI advancements, focusing on their practical applications. This episode covers several major developments in the AI landscape:

Eleven Labs' Innovation in Audio AI

• Eleven Labs has launched an AI Podcast Generator through their 11 Reader iOS app, enabling podcast creation from various text sources in 32 languages.
• The company has also introduced a platform for building custom AI agents with configurable voices and response styles.

Amazon and OpenAI's New Models

• Amazon has introduced Nova, their foundational model focused on math, science, coding, and reasoning tasks.
• OpenAI has launched a $200/month ChatGPT Pro tier, providing advanced access to GPT-4's capabilities.

Microsoft and Meta's Developments

• Microsoft's Copilot Vision enables screen-aware AI assistance within the Edge browser.
• Meta's Llama 3 demonstrates improved efficiency through quality training data.

Productivity and On-Device AI

• The Twos app introduces PAL (Personal Active List) for AI-powered task management.
• Apollo AI brings on-device AI capabilities to iOS devices.

Google's Advances

• Gemini exp-1206 features a 2 million token context window, surpassing ChatGPT 4.0 on LM Arena.
• The Illuminate experiment enables podcast creation with customizable styles.

Breakthrough Technologies

• OpenAI's Sora introduces advanced text-to-video generation capabilities.
• Google's Willow quantum computing chip achieves significant computational breakthroughs.
• Sundar Pichai proposes space-based quantum computing collaboration with Elon Musk.

Links:

• Apollo AI
• Copilot Vision
• ElevenLabs
• Google Illuminate
• ElevenReader (iOS and Android)
• OpenAI Sora
• Twos

In this episode of "AI Wanta Talk," host Jacob Norgord explores recent AI advancements, focusing on their practical applications. This episode focuses on four major developments in the AI landscape:

Anthropic AI's Claude New Features

• Anthropic AI has introduced two new features for its Claude AI chatbot: Styles and Model Context Protocol (MCP).
• Styles enables users to customize how Claude responds using presets like "concise," "explanatory," or "formal."
• Model Context Protocol (MCP) acts as a "universal translator" for AI and data sources, allowing Claude to connect with external sources like files or websites and interact with them.
• MCP enables Claude to perform complex tasks such as generating images based on user requests, writing code, and integrating images into websites.

Microsoft AI's Apparent Long-Term Memory Breakthrough

• Microsoft AI CEO Mustafa Suleyman believes long-term memory is the crucial missing element in current AI chatbots.
• Microsoft AI is working on incorporating long-term memory into its Copilot chatbot, enabling it to retain information from previous conversations and use it to provide more personalized and accurate responses.
• The goal is to eliminate the need for users to constantly re-explain context and make interactions with AI more natural and efficient.

Bronze AI Acquires Chroma

• Bronze AI, known for its innovative Bronze file format that creates dynamic music experiences, has acquired Chroma, a company specializing in audiovisual entertainment for mobile devices.
• The acquisition suggests potential for combining Bronze's evolving music with Chroma's visual expertise, creating a multisensory experience for listeners.
• The collaboration may push the boundaries of art by blending music and visuals in innovative ways.

Links:

• More info on Claude Styles
• More info on Model Context Protocal (MCP)
• Pi
• Mustafa Suleyman Interview Excerpt
• Dot by New Computer
• Bronze
  • “Jasmine” by Jai Paul (Bronze Version)