- The Rundown AI
- Posts
- Meta reveals 'Orion' glasses!
Sign Up | Advertise | Tools | AI University
Welcome, AI enthusiasts.
Meta just unveiled the world’s most advanced AR x AI glasses alongside major AI updates, including Llama 3.2 and a new Voice Mode for its 500M Meta AI users.
Is Zuck’s vision of an AI-powered metaverse finally becoming a reality? Let’s get into it…
In today’s AI rundown:
Meta unveils AR x AI glasses, new models, and more
OpenAI CTO exits amid rumors of non-profit removal
Analyze images with AI in seconds
AI breakthrough in treating rare diseases
5 new AI tools & 4 new AI jobs
More AI & tech news
Read time: 4 minutes
LATEST DEVELOPMENTS
META
Image source: Reality Labs
The Rundown: At its Connect 2024 conference, Meta revealed a host of new AI announcements, including its new Orion AR x AI glasses, Llama 3.2, AI features for Reels, and major updates to Meta AI—including a new Voice mode.
The details:
Orion AR glasses prototype boasts a sub-100g weight, wide field of view displays, and advanced features like voice control and hand tracking, taking Meta over 10 years to build.
Meta introduced Llama 3.2, its first major vision model capable of understanding both images and text, with 11B and 90B parameter versions.
New super-small 1B and 3B parameter Llama models were also announced, optimized for on-device use in smartphones and potentially future glasses.
New AI features are coming to Instagram, including automatic video dubbing and lip-syncing for creators for any language and AI-generated content, ‘Imagined for you‘ on Feeds.
Meta announced Voice Mode, similar to ChatGPT’s recent Advanced Voice Mode, which allows users to use their voice to talk with Meta AI on Messenger, Facebook, WhatsApp and Instagram DMs.
Why it matters: It’s difficult to overstate the significance of Meta Connect 2024. With new open-source models, the most advanced AR glasses ever made, and nearly 500 million monthly active Meta AI users now getting AI Voice chat directly onto their favorite platforms—the tech giant is showing, once again, never bet against Zuck.
TOGETHER WITH DEFINED.AI
The Rundown: Defined.ai’s new Data Access Plan (DAP), offers high-quality AI speech training data at a significantly reduced cost – slashing costs by up to 55% compared to standard rates.
The Data Access Plan offers you:
Responsible, high-quality speech data across numerous languages and locales for all your fine-tuning and training needs
Customizable data plans that scale as your business grows
Instant access to off-the-shelf datasets curated by a Defined.ai expert
Join Defined.ai’s webinar on October 10th to learn more.
OPENAI
Image source: Fortune
The Rundown: Mira Murati, (previously) CTO at OpenAI, just announced her decision to leave the company after six and a half years amid rumors that the company is removing its non-profit control and giving CEO Sam Altman equity.
The details:
Murati is apparently leaving to create time and space for her own exploration, while focusing on ensuring a smooth transition.
This transition aligns with a recent report that OpenAI will restructure its core business into a for-profit benefit corporation with Sam Altman and the non-profit part of the company getting a minority equity position.
Murati expressed deep gratitude for her time at OpenAI, highlighting recent achievements like speech-to-speech technology and OpenAI o1.
CEO Sam Altman responded with appreciation for Murati’s contributions, saying the company will ‘soon’ announce transition plans.
Why it matters: This is another big shakeup for OpenAI after losing Andrej Karpathy in February, Ilya Sutskever in May, and Greg Brockman (who’s on a sabbatical) in August. With Sam Altman as the last high-profile leader left, the world has many questions—but has only been answered with the classic OpenAI-style ‘soon‘.
AI TRAINING
The Rundown: GroqCloud's new LLaVA v1.5 7B model allows users to analyze images and get AI-powered insights with near-instantaneous results.
Step-by-step:
Visit the GroqCloud Developer Console.
In the top-right corner, open the models dropdown menu and select "llava-v1.5-7b-4096-preview".
Upload your image by dropping it into the designated area or using the file browser.
Enter a prompt or question about the image in the "MESSAGE" field below.
Click "Submit" to generate AI analysis of your image.
PRESENTED BY DIGITALOCEAN
The Rundown: DigitalOcean’s cutting-edge platform empowers developers with scalable GPU solutions and user-friendly tools to help you bring your AI/ML apps to life quickly and affordably.
DigitalOcean offers:
Accessible GPU solutions to deploy, test, and improve AI applications
A comprehensive cloud computing platform tailored for seamless AI/ML development
Upcoming GPU Droplets powered by NVIDIA H100 GPUs for advanced AI training and deep learning
AI RESEARCH
Image source: Midjourney
The Rundown: Harvard Medical School researchers recently developed an AI model called TxGNN that can identify existing drugs for repurposing to treat rare and neglected diseases.
The details:
TxGNN identified drug candidates from nearly 8,000 existing medicines for over 17,000 diseases, many without current treatments.
The model outperformed leading AI drug repurposing tools by nearly 50% in identifying candidates and was 35% more accurate in predicting contraindications, a specific situation in which a medcine should not be used.
TxGNN uses a novel approach that identifies shared features across multiple diseases, allowing it to extrapolate from well-understood conditions to poorly understood ones.
The researchers have made the tool freely available to encourage its use by clinician-scientists in the search for new therapies, especially for rare and untreated conditions.
Why it matters: Another week, another insane medical breakthrough for AI. While we still need years of clinical validation and approvals before widespread use, TxGNN has the potential to save thousands of lives and improve the lives of people who likely thought a treatment for their specific disease would never come.
NEW TOOLS & JOBS
🚀 Notion AI - Search and chat with documents across Notion, Slack, and Google Drive
📊 Rows AI Analyst 3.0 - An AI data analyst that visualizes and formats data
🖼️ Magnific Mystic V2 - Advanced AI generator that can output up to 4k resolution images
💡 Magic Patterns - Generate product design and React code
🎵 OpenMusic - Create custom tunes from text descriptions
🖥️ Anyscale - Software Engineer
⚡ Waymo - Charging Infrastructure Program Manager
🧪 Mistral AI - Research Engineer
🩺 Curai - Physician (Telemedicine)
QUICK HITS
OpenAI is reportedly developing an improved version of its Sora AI video generation model, aiming for higher quality and longer clips than previously demonstrated.
Meta announced it will not immediately join the European Union’s voluntary AI Pact, instead focusing on compliance with the upcoming AI Act regulations.
Nvidia analysts predicted the company will produce around 450,000 Blackwell AI GPUs in Q4 2024, potentially generating over $10 billion in revenue despite initial production challenges.
Nebius Group revealed plans to invest over $1 billion in AI infrastructure across Europe by mid-2025, including GPU clusters and data centers.
The Federal Trade Commission announced enforcement actions against multiple companies for deceptive or unfair use of artificial intelligence in their practices.
OpenAI CEO Sam Altman said the Advanced Voice Mode rollout for ChatGPT has been completed early, except in jurisdictions requiring additional external review.
THAT’S A WRAP
SPONSOR US
Get your product in front of over 650k+ AI enthusiasts
Our newsletter is read by thousands of tech executives, investors, engineers, managers, and business owners around the world. Get in touch today.
FEEDBACK
How would you rate today's newsletter?Vote below to help us improve the newsletter for you. |
If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.
Reply