• AI Emergence
  • Posts
  • Meta ♥️ AI, AR, VR: All comes alive at Meta Connect 2024

Meta ♥️ AI, AR, VR: All comes alive at Meta Connect 2024

Along with: Mira Murati steps down as OpenAI’s CTO, followed by two senior leaders

Hey there, 

What an incredible week!

Meta set the stage on fire with Orion, Llama 3.2, Quest 3s, and a transparent RayBan (I love it). On the other side - there is something big brewing inside OpenAI. Google dropped the price on its models and Jony Ive is cooking up a new device!

So much packed - all under a week!

Now coming to this week’s newsletter…

What would be the format? Every week, we will break the newsletter into the following sections:

  • The Input - All about recent developments in AI

  • The Tools - Interesting finds and launches

  • The Algorithm - Resources for learning

  • The Output - Our reflection 

Table of Contents

Meta has announced “Orion,” potentially the first consumer-grade, full holographic AR glasses. They’re lightweight and equipped with hand-tracking and eye-tracking – a true neural interface. Watch the video to see them in action.

Meta has also launched the latest versions of its AR headsets: Quest 3s and Quest 3.

Ray-Ban glasses are getting a major upgrade: real-time AI video processing, allowing you to ask the glasses questions about what’s in front of you, live language translation, and integration with music streaming apps.

Meta has released Llama 3.2, featuring small and medium vision models (11B and 90B) along with lightweight text-only models (1B and 3B) optimized for mobile and edge devices.

Meta has introduced new multimodal features:

  • Meta AI gets a voice.

  • You can now share photos with Meta AI, ask questions, or even edit them on Instagram and Facebook.

Currently in testing: Automatic language translation in reels, dubbing, and lip-syncing. Meta AI simulates the speaker’s voice in another language and syncs their lips to match (source)

Murati’s departure, alongside two other senior staffers, comes as the company is preparing to announce a new structure that will see its for-profit arm no longer subservient to the board of its nonprofit foundation.

​​“I’m stepping away because I want to create the time and space to do my own exploration,” she said in her statement on X.

She played a pivotal role in the company's AI advancements and shared her decision to focus on personal exploration through a message on X. 

Altman also confirmed the resignations of senior leaders Barret Zoph (VP of Research) and Bob McGrew (Chief Research Officer). 

To maintain continuity, Altman appointed Matt Knight as Chief Information Security Officer and Josh Achiam as Head of Mission Alignment. (source)

Google has introduced two of its Gemini 1.5 models, Gemini 1.5 Pro 002 and Gemini 1.5 Flash 002, and that too within a month of releasing its previous version.

These models are said to be improved performance-wise with higher output, lower cost, and better adherence to user instructions through updated filters, which have been enhanced to better follow prompts while keeping safety measures in place. 

If compared to the previous model, it shows a boost of up to 20% in MATH and hiddenMath Benchmarks. 

Additionally, rate limits for users have been increased to 2,000 requests per minute for the Flash model and 1,000 for the Pro model, alongside faster token generation for long text blocks. 

Accessible for developers for free via Google AI Studio and the Gemini API and to enterprises via Vertex AI. (source)

At Capital Markets Day 2024, T-Mobile CEO Mike Sievert and OpenAI’s Sam Altman discussed the evolving AI landscape and its implications for various industries.

Altman highlighted OpenAI's recent introduction of the o1 family of "reasoning" models, marking a significant leap from their GPT series by focusing on advanced reasoning capabilities.

He outlined a five-level progression of AI capabilities, noting that the shift from chatbots to reasoning systems will pave the way for autonomous AI agents.

Their collaboration led to the development of IntentCX, an AI-driven customer engagement platform that enhances support by understanding and anticipating customer needs.

Altman also envisioned AI's transformative potential in healthcare and education, while acknowledging the challenges of data privacy and ethical development. (source)

OpenAI now offers an advanced voice feature in ChatGPT, enabling more natural and fluid audio conversations. 

This update is currently available to premium users on Plus, Team, or Enterprise plans, with more regions gaining access over time. 

Users can select from nine voices, personalize responses, and experience improved speed and accents in various languages. However, advanced voice mode has daily usage limits for these plans.. (source)

NVIDIA has launched AI Aerial, a comprehensive suite of accelerated computing software and hardware designed for AI radio access network (AI-RAN) technology, which will enable network optimization at scale, leading to cost savings and new revenue streams. 

AI Aerial supports diverse applications, including teleoperations for robots and vehicles, computer vision, and generative AI co-pilots. The platform provides tools for high-performance, software-defined RAN, and features like the Aerial Omniverse Digital Twin for simulating wireless systems. (source)

Snap has just introduced an upgraded version of its spectacles augmented-reality glasses, signaling its continued focus on wearable devices as the next frontier in technology.

The fifth-generation Spectacles, feature improved AR capabilities, responding to hand gestures and voice commands, while also offering a larger field of view and automatic sun-tinting.

However, the new glasses will initially only be available to developers for $99 per month, with the goal of enhancing AR experiences and driving consumer adoption. (source)

Jony Ive, Apple's renowned designer, is collaborating with OpenAI CEO Sam Altman to develop an AI-driven product. 

While specific details remain undisclosed, strong investor interest could lead to the project raising $1 billion by year-end. 

Ive's design company, LoveFrom, will oversee the project, which has attracted notable investors such as Emerson Collective. 

The partnership between Ive and Altman originated from a dinner arranged by Airbnb CEO Brian Chesky, where they explored the potential of generative AI in future computing devices. (source)

ImagetoCaption.ai is an AI-powered tool designed to generate accurate image captions automatically. Using advanced image recognition technology, it identifies objects, scenes, and relevant details within a photo to create descriptive captions, making it ideal for accessibility, social media content, or digital marketing.

To access Imagetocaption.ai :

  1. Visit imagetocaption.ai.

  2. Upload the image you want to caption.

  3. Select the theme, location, prompt, and details you want to have in the caption.

  4. The AI processes the image and generates a descriptive caption based on its content.

  5. You can then use or refine the caption for your needs.

This is what I generated from my trial of the tool:

Image:

Caption Generated: Unlock the secrets of real-world RAG systems with our free course! Join Dipanjan Sarkar and elevate your AI expertise. 🚀 Don’t miss this opportunity—sign up today!

#AnalyticsVidhya #AI #DataScience #Community #FreeCourse

  • In the recent episode of Leading With Data, I had an engaging conversation with Rohan Rao, Principal Data Scientist at H2O.ai. He discussed his views on how he approaches selecting the right LLMs for different business needs and why he believes that understanding the trade-offs between proprietary and open-source LLMs is crucial for making informed decisions.

  • Sam Altman recently shared his blog titled "The Intelligence Age." where he discussed how AI advancements, particularly through deep learning, have unlocked unprecedented opportunities for human progress.

I am trying out Apple Intelligence this week in Developer Beta! Have you tried it - what has been your experience? 

Look forward to hearing from you.

How do you rate this issue of AI Emergence?

Would love to hear your thoughts

Login or Subscribe to participate in polls.

Reply

or to participate.