- AI Emergence
- Posts
- Now Build Applications in a Breeze! Future of AI-Generated Videos
Now Build Applications in a Breeze! Future of AI-Generated Videos
Along with - AI 'Godfather' Hinton and Princeton's Hopfield Win Nobel Prize
Analytics Vidhya (Curated by Kunal Jain)
October 10, 2024
Hey there,
A big congratulations to all the Nobel prize winners in different fields. It is fascinating and inspiring to see AI take center stage to win the Nobel Prize in Physics and Chemistry - who would have thought?
But that makes me think, what would be the future of Nobel prizes in a decade? How would we divide the credit between the researcher and the AI assisting the researcher? Would we have a Nobel prize separately for AI or would AI be ineligible to participate in Nobel prizes?
Something to ponder on!
In the meanwhile Meta announced Movie Gen and some of the videos I have seen are amazing - waiting for its public release and testing before forming an opinion on the same.
Let’s go through the developments of this week
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
The 2024 Nobel Prizes in Physics and Chemistry highlight AI's pivotal role in modern scientific breakthroughs.
In physics, John Hopfield and Geoffrey Hinton's (known as the “godfather of AI”) work on neural networks revolutionized how AI learns and processes information, laying the foundation for current AI systems. Their blend of neuroscience and physics has enabled machines to recognize patterns and “think” like humans.
In chemistry, Demis Hassabis, John Jumper, and David Baker used AI to solve the decades-old mystery of protein folding and design new proteins.
These advancements demonstrate AI's growing influence in reshaping science, from biology to physics. (source)
Meta has announced its AI-powered video generator, Movie Gen, known for its high-definition videos with sound from text inputs.
It allows users to generate new clips, edit existing videos, and enhance still images with AI-generated audio and video elements. Moreover, it can also customize videos by altering visual components and adding AI-generated sounds like ambient noise and music.
While the tool shows promise, Meta is not ready to launch it publicly due to high costs and long processing times.
This follows similar efforts by competitors like OpenAI, though AI video generators continue to raise concerns over copyright, data usage, and the future of creatives. Meta stated it trained Movie Gen using licensed and publicly available datasets but didn’t provide specifics. (source)
ChatGPT is introducing Canvas as its new interface, designed for more complex writing and coding projects beyond simple chats.
Canvas lets users collaborate with ChatGPT on projects by refining and editing ideas in real-time, providing inline feedback, and making targeted edits. It also improves collaboration by highlighting specific sections, simplifying revisions, and offering shortcuts for writing and coding tasks.
This feature is available in beta for GPT-4o users and will soon be available to all users. (source)
Hugging Face released the OpenAI-Gradio package, a Python tool that lets developers create AI-powered web apps with minimal code.
The package reduces development time and makes AI more accessible to businesses of all sizes by simplifying the integration of OpenAI's LLMs.
Developers can easily install the package, set their OpenAI API key, and launch custom AI solutions using the Gradio interface. (source)
The recently launched iPhone 16 lacks Apple Intelligence, but that’s about to change with the release of iOS 18.1 on October 28th. Moreover, Apple Intelligence will soon be available on iPhone 15 Pro and newer models, as well as Macs and iPads with M1 chips.
The update will introduce AI-powered notification summaries, along with improvements to Siri's ability to understand speech, and visual changes like a rainbow-colored glowing edge when summoning Siri.
While some features, such as Apple’s Image Playground and Genmoji, are expected by December, Siri’s ability to perform tasks across apps may not arrive until March. (source))
Imagine wearing glasses that instantly display a person’s name, address, phone number, and even details of relatives every time you see them. Sounds wild, right? Two hackers have demonstrated that this is already possible!
Harvard researchers AnhPhu Nguyen and Caine Ardayfio have developed a pair of AI-powered smart glasses, dubbed I-XRAY, capable of doxxing individuals by extracting personal information just by looking at them.
The glasses combine off-the-shelf facial recognition technology, public databases, and large language models to gather real-time data.
The researchers have claimed that this is only to raise awareness about the privacy risks of advanced AI and smart tech. (source)
At the AI Summit in Washington, D.C., Nvidia showcased its latest technology and announced seven key initiatives to advance AI adoption across various industries.
The company collaborates with U.S. tech leaders, including Accenture and Deloitte, to help organizations develop custom AI applications using Nvidia’s NIM Agent Blueprints and NeMo microservices.
Nvidia’s new NeMo microservices, such as NeMo Customizer and NeMo Evaluator, enable developers to curate data, customize models, and deploy solutions seamlessly across various platforms. (source)
Dr. Robert Yang, former MIT assistant professor is to launch Altera, a research lab focused on developing "digital humans" that can interact and collaborate with people in meaningful ways.
Alongside his co-founders, Yang used OpenAI’s GPT-4 models to create autonomous AI agents that simulate human cognitive and emotional responses, allowing these agents to play games like Minecraft and form long-term relationships with users.
Altera addresses the challenge of data degradation in AI, enhancing the agents' ability to operate autonomously for extended periods. (source)
Tool: OpenAI Canvas
ChatGPT is introducing Canvas as its new interface, It allows users to interact with generative AI models in real-time by offering a flexible, visual workspace where they can brainstorm, design, and build ideas.
Problem statement: I built a simple Snake game using ChatGPT 4 on Canvas. You can easily highlight parts of the code and ask ChatGPT questions. It also offers cool features like commenting, bug fixing, and porting to different languages- a fantastic toolset for all coders!
How to access:
Sign in to OpenAI
Click on ChatGPT-4 with Canvas
Enter your prompt
Get your response instantly
Click here to view my trial of the tool.
For those interested in overcoming practical challenges in building Retrieval-Augmented Generation (RAG) systems, this course by Analytics Vidhya offers solutions to optimize data retrieval, minimize hallucinations, and improve system performance through advanced techniques like embedding models and context selection.
If you're curious about how AI is impacting businesses, this blog by Ethan Mollick offers insights into current challenges and practical strategies. It explores how AI can boost productivity and why companies need to invest in their own AI R&D. A must-read for anyone navigating AI integration in the workplace.
In this video conversation, Lex Fridman and the Cursor team discuss how artificial intelligence is revolutionizing software development, improving user experience, and boosting productivity. The conversation also touches on the broader implications of AI, including its societal impact, ethical considerations, and user privacy.
For those interested in optimizing LLMs for their business needs, this course by Analytics Vidhya offers a comprehensive framework to choose the right LLM, considering key factors such as accuracy, cost, and real-world application.
If you're looking to build a strong foundation in Agentic AI systems, this course covers the history, current applications, and multi-agent system dynamics. It offers a practical guide to understanding and implementing intelligent agent systems.
That’s it for this week!
What do you think about AI participation in Nobel Prizes? It is not something we would probably come across in the near future, but once it happens - there will be no going back!
Would love to hear your thoughts on it!
How do you rate this issue of AI Emergence?Would love to hear your thoughts |
Reply