
Create articles from any YouTube video or use our API to get YouTube transcriptions
Start for freeMeta Launches New AI App
Meta held their first Llamacon event this week, focused specifically on AI updates. The biggest announcement was a rebranding and expansion of their Meta View app, now called the Meta AI app.
Key features of the new Meta AI app include:
- Standalone AI chat feature powered by the Llama 4 language model
- Ability to generate images using Meta's EMU image generator
- Social sharing of AI conversations and creations
- Continuity between conversations on Ray-Ban Meta smart glasses and the mobile/web app
Meta also updated their privacy policy for Ray-Ban Meta glasses:
- AI with camera use is always enabled unless manually turned off
- Photos/videos captured are not used for AI training
- Voice recording storage can no longer be disabled, but recordings can be deleted
- Voice data may be used to improve Meta's language models
Mark Zuckerberg mentioned plans to eventually incorporate ads or product recommendations into Meta AI, likely after scaling usage for about a year.
Google Expands AI Search Mode
Google is rolling out access to their new AI-powered search mode:
- Now available to all US Labs users
- Small percentage of general US users will see it in coming weeks
- Provides AI-generated responses with links, similar to Perplexity or ChatGPT
Other Google AI updates:
- Gemini app on Android can now edit uploaded images
- NotebookLM can generate audio overviews in 50+ languages
- New "Little Language Lessons" app for quick language learning
Recraft Launches Expanded Style Library
Recraft, an AI image generation platform, released a major update to their style system:
- Huge new library of pre-made styles to choose from
- Ability to save and customize styles
- Option to blend multiple styles together
- Styles can be shared via link
This allows users to rapidly iterate on styles and maintain brand consistency across generated images.
OpenAI Reverts GPT-4 Personality Changes
OpenAI rolled back recent updates to GPT-4 after feedback that it had become overly flattering and agreeable. They are revising how they incorporate user feedback to focus more on long-term satisfaction.
Other ChatGPT updates:
- Improved web search capabilities
- New shopping experience with product carousels
- Search functionality added to WhatsApp integration
- Trending topics and autocomplete in prompt window
Elon Musk Teases Grok 3.5
Elon Musk announced Grok 3.5 will be available next week for top-tier X Premium+ subscribers. He claims it will be the first AI that can accurately answer technical questions about topics like rocket engines by reasoning from first principles.
Anthropic Announces Claude Integrations
Anthropic is adding official integrations to allow Claude to connect with other apps via Model Context Protocols (MCPs). This feature is currently limited to higher-tier plans.
Alibaba Unveils Qwen 3 Hybrid Reasoning Model
Alibaba released Qwen 3, a new large language model that allows toggling of step-by-step reasoning. This "hybrid reasoning" approach aims to balance speed and accuracy.
Vercel Introduces Vi AI Computer Assistant
Vercel announced Vi, an AI agent that can interact with applications on your computer. It aims to perform tasks by understanding and manipulating user interfaces. The tool is available for download but access is currently waitlisted.
Midjourney Adds Omni Reference Feature
Midjourney version 7 now includes an "omni reference" capability, allowing users to inject specific elements from reference images into generated scenes.
Krea AI Launches Instant Film Effect
Krea AI released a new feature that converts portrait images into animated, Polaroid-style videos.
Higgsfield AI Introduces Iconic Scenes
Higgsfield AI debuted a tool that places user selfies into animated recreations of famous movie scenes.
Suno Releases Version 4.5
AI music generator Suno updated to version 4.5, with improvements to genre variety, voice quality, and prompt adherence.
Companies Embracing "AI-First" Approaches
Following Shopify's lead, Duolingo announced a shift to become an "AI-first" company. They plan to gradually replace contractor work with AI solutions where possible, aiming to free up employees for more creative tasks.
Lyft Launches AI Earnings Assistant
Lyft introduced an AI tool to help drivers optimize their shifts based on real-time data about airport arrivals, local events, and other factors affecting ride demand.
Aurora Deploys Driverless Trucks in Texas
Autonomous vehicle company Aurora has begun operating fully driverless semi trucks on public highways between Dallas and Houston, having completed over 1,200 miles of autonomous driving.
Conclusion
This week's AI news highlights the rapid pace of development across various sectors. From Meta's push into consumer AI apps to advancements in autonomous vehicles, we're seeing AI capabilities expand and integrate more deeply into daily life and business operations. As large language models become more refined, much of the exciting innovation is happening in creative AI tools and practical applications that leverage these underlying technologies in novel ways.
Article created from: https://youtu.be/w20FT5quE_k?feature=shared