- AI Emergence
- Posts
- OpenAI Launches o1 pro as Part of ChatGPT Pro Subscription!
OpenAI Launches o1 pro as Part of ChatGPT Pro Subscription!
Along with: Amazon goes all out to take the competition head-on!
Analytics Vidhya (Curated by Kunal Jain)
December 06, 2024
Hey there,
What an incredible week!
Makes me think - we as humans and organisations tend to slow down in holiday months. Rarely do you see large product announcements in Big Tech in December.
Well - AI is different! It seems like no one in the Big Tech is slowing down this year. If anything - there is a ton of action happening. OpenAI launched o1 pro, which on benchmarks, is performing better than o1. It is going to make announcements for the next 11 days. Amazon, on the other hand, launched a suite of products, chips, and models at eye-popping costs. Google launched its own models.
Let’s look at all the action which happened 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
Big tech companies usually go all out with flashy launches, but OpenAI is taking a unique approach with a 12-day streak of launches and demos showcasing their latest capabilities. Here’s what went down on Day 1:
OpenAI introduced o1 for ChatGPT Plus users and launched o1 Pro as part of their new ChatGPT Pro plan, which costs a whopping $200/month—almost like adding an extra team member!
The o1 model now supports multimodal capabilities, so you can interact with both text and images. It’s also faster and smarter, with a 34% reduction in errors on tough problems compared to its preview version. The response time is quicker too, making interactions smoother and more efficient.
For o1 Pro, OpenAI claims that it is much more reliable. The model is only considered successful if it solves a question correctly 4 out of 4 times.
The new ChatGPT Pro subscription gives you unlimited access to models like GPT-4o, o1 preview, and their top-tier ultimate model.
OpenAI also released the o1 Systems Card, sharing how they ensured safety before launching o1 and o1-mini.
We’re excited to see what’s next—stay tuned for more updates! (source)
AWS re:Invent 2024 is underway in Las Vegas, showcasing cutting-edge innovations in cloud computing and AI. Here’s a snapshot of the big updates, with some quick context to help you understand why they matter:
Amazon Bedrock Updates
Amazon Bedrock simplifies building and running generative AI models without heavy infrastructure. Key upgrades include:
Multi-Agent Orchestration: Enables AI agents to collaborate, improving workflows. Moody’s uses it for accurate financial analysis.
Automated Reasoning Checks: Reduces errors like “AI hallucinations” for more reliable outputs.
Model Distillation: Builds smaller, faster AI models for efficient scaling.
Amazon SageMaker Enhancements
SageMaker is AWS’s go-to platform for building and deploying machine learning models. New features include:
Unified Data and AI Hub: Merges data and AI tools into one platform for faster development.
HyperPod Task Governance: Optimizes GPU usage, cutting idle time and costs by 40%.
Amazon Nova Models
Nova is AWS’s new family of foundation models for generative AI:
Understanding Models: Handle text, images, and videos to generate outputs for tasks like summarization and content creation.
Creative Content Models: Generate high-quality images (Nova Canvas) and videos (Nova Reel), perfect for marketing and media.
With these updates, AWS is making AI smarter, faster, and more accessible. Stay tuned for more highlights from re:Invent 2024! (source)
OpenAI is considering advertising as a potential revenue stream, although CFO Sarah Friar has cautiously stated that there are "no active plans" at the moment.
Friar mentioned in an interview with the Financial Times that they are exploring an ads-based business model with plans to be "thoughtful" about where and how ads are placed. However, she emphasized that no active steps have been taken in this direction. Notably, this idea is under consideration, highlighted by the recent hiring of ex-Google executive Shivakumar Venkataraman.
Despite these developments, the high costs of AI development make such a shift appear increasingly likely.
This potential move contrasts with founder Sam Altman's recent comments at Harvard, where he described the combination of AI and advertising as "uniquely unsettling" and a "last resort." (source)
Google DeepMind's GenCast represents a remarkable leap in weather forecasting technology. This AI model outperforms current leading systems, including the prestigious European Centre's ENS model, particularly in predicting extreme weather events.
Instead of using traditional physics-based calculations that take hours to process, GenCast analyzes historical weather data to understand complex atmospheric patterns, delivering forecasts up to 15 days ahead in just minutes.
This breakthrough suggests we're entering a transformative era where AI-powered weather prediction could offer unprecedented speed and reliability, potentially revolutionizing how we anticipate and respond to weather events. (source)
DeepMind’s Genie 2 is taking AI creativity to the next level, enabling the generation of infinite, interactive 3D environments from a single image prompt.
What makes it special? Unlike its 2D predecessor, Genie 2 introduces advanced features like:
Physics simulation: Realistic movements and interactions.
Character animation: Lifelike actions and gestures.
Object interaction: Dynamic and responsive environments.
Persistent memory: Off-screen elements remain intact, adding depth to experiences.
Teamed up with Imagen 3, a text-to-image model, Genie 2 lets users describe a world in words and then step into it. Imagine prototyping interactive experiences or training embodied AI in virtual worlds that respond to both human and AI inputs—an exciting leap forward! (source)
World Labs, founded by Fei-Fei Li, has introduced an innovative AI system capable of generating interactive 3D scenes from a single image.
Unlike traditional AI tools, these scenes are modifiable, physics-aware- maintaining stability, solidity, and depth- and offer consistent quality with features like dynamic lighting and object interaction.
The system renders scenes live in-browser, complete with a controllable camera and an adjustable depth of field for creating background blur effects. It’s like taking a picture of a room; with the tool, you can step into the picture on your screen, look around the room from different angles, adjust the lighting, and more.
Though still in its early stages, this technology promises to revolutionize industries such as gaming, movies, and simulation by providing a cost-effective way to create immersive 3D environments. (source)
Elon Musk has filed a preliminary injunction to stop OpenAI’s shift to a for-profit structure, arguing it threatens the company’s original nonprofit mission. The lawsuit also targets Microsoft, CEO Sam Altman, and former board members.
The filing accuses them of anticompetitive practices, sharing sensitive data with Microsoft, and self-serving business dealings. Musk also highlights Altman’s financial ties to companies like Stripe, which provides services to OpenAI.
OpenAI has called the claims baseless, but this comes as the company races toward a $150 billion valuation, a move that could shape its future and impact the broader AI industry. (source)
Orbital Materials has partnered with AWS to develop AI-driven technologies for data center decarbonization and efficiency.
Orbital’s proprietary AI platform accelerates the design, synthesis, and testing of advanced materials, achieving a 10x improvement in carbon removal performance since early 2024.
Their collaboration includes deploying Orbital’s open-source AI model, Orb, on AWS platforms, enabling faster R&D for semiconductors, batteries, and electronics.
AWS will support Orbital with Amazon SageMaker HyperPod and Trainium for large-scale training and cost efficiency. Orbital plans to test its carbon removal technology by 2025, aiming to revolutionize Data Center sustainability and promote broader innovation in climate technology. (source)
Intel CEO Pat Gelsinger has stepped down less than four years into his tenure after the board expressed dissatisfaction with his turnaround strategy's progress.
Gelsinger's ambitious plans, including a $20 billion investment in new factories and a shift to contract chip manufacturing, failed to yield results amid declining PC demand and competitive pressure from Taiwan Semiconductor and Nvidia.
Under his leadership, Intel's stock plummeted over 60%, prompting the board to appoint CFO David Zinsner and executive Michelle Johnston Holthaus as interim co-CEOs. The company, receiving $7.86 billion in U.S. subsidies, now seeks a new leader to restore investor confidence. (source)
A coalition of Canada’s leading news organizations, including CBC, The Globe and Mail, and Toronto Star, has filed a lawsuit against OpenAI, alleging the illegal use of news articles to train its ChatGPT software.
Filed in Ontario’s Superior Court, the suit seeks billions in damages, an injunction to stop OpenAI from using articles without consent, and disgorgement of profits.
This marks the first united legal action by a country’s major publishers against OpenAI, citing unauthorized scraping of proprietary content.
OpenAI, valued at $157 billion, faces increasing global litigation over similar claims, despite having licensing agreements with some media organizations. (source)
Tool: FoodiePrep
FoodiePrep creates tailored recipes and meal plans to fit your preferences and lifestyle. From meal planning to recipe creation, FoodiePrep makes dining effortless and enjoyable!
How to access it:
Login
Chat with Ask Foodie
Share Allergies/Dietary Needs
Explore Recipes
Organize Your Recipes
Generate Shopping Lists
I tried this and here's what I got in suggestion from the tool.
In the recent episode of Leading with Data, I had a conversation with Alan Nichol of Rasa about his experience and views on conversational AI, its evolution, and emerging trends in the field.
This interview between Ann Bordetsky of NEA and Aravind Srinivas, CEO of Perplexity AI, discusses how Perplexity is transforming search technology by integrating contextual search with verified, source-based answers.
If you're exploring the fast-evolving AI landscape, AI Engineering by Chip Huyen is a must-read. It offers practical insights into using foundation models with clear frameworks and deployment strategies. It is available now in Kindle format.
The course "A Comprehensive Learning Path to Become a Data Analyst in 2025" by Analytics Vidhya, offers a structured guide for beginners and career shifters to learn data analytics, featuring in-demand 2025 skills, curated resources, and hands-on exercises for a successful career.
What do you think about ChapGPT Pro? Would you shell out $200 monthly for O1 Pro and unlimited usage? It is still cheaper compared to a human of similar capability, but then there are a ton of lower-cost AI available that might do most of the tasks.
Are you going to subscribe to ChatGPT Pro? |
How do you rate this issue of AI Emergence?Would love to hear your thoughts |
Reply