The ChatGPT Observer Edition 28

If you're short on time, here's your one-minute summary of this week's key AI news and advancements:

?? Top Story: OpenAI launches O1, a new AI model excelling in reasoning for complex STEM problems, available in two variants: O1 and O1-mini.

?? Key Statistics: 57% of enterprises use Generative AI, with 82% viewing it as a top priority for their AI strategy.

?? This Week's Spotlight: Apple introduces Apple Intelligence, a new AI-powered system built into iOS 18.

?? ChatGPT & OpenAI News: ChatGPT reaches over 1 million business users; OpenAI invests in a drug discovery startup; ChatGPT experiences a global outage on September 11, 2024.

? Claude Sonnet News: Claude AI updates with code highlighting and editing features.

?? Corporate AI Developments: Google's NotebookLM adds voice explanations for complex topics; Mistral introduces 'PixTral', its first multimodal AI model.

?? AI Innovations: Replit launches as an online collaborative coding platform.

?? AI in Business: Connectly uses AI for personalized customer messaging; Bill Gates leverages AI for meeting summaries.

?? AI in Education: Australia prioritizes integrating generative AI into schools.

? Responsible AI: Australian government plans to outlaw doxxing; U.S. Senate leaders urge FTC to investigate AI content summaries; Meta faces scrutiny over using Australian content to train AI models.

?? AI Learning: New resources on AI prompt engineering, Retrieval-Augmented Generation (RAG), and Elon Musk's vision of AI's future impact.

?? For more details, read the full newsletter!

?? Top Story: Introducing OpenAI o1

OpenAI has launched its newest AI model, O1, previously codenamed "Strawberry." Designed to excel in reasoning, O1 is built to tackle complex problems in science, coding, and mathematics by spending more time "thinking" before generating responses, mimicking human-like thought processes.

Two Variants of O1:

• OpenAI O1: A powerful model tailored for advanced reasoning in STEM fields. It features robust safety and alignment measures, making it ideal for high-stakes problem-solving.
• OpenAI O1-mini: A smaller, faster, and 80% cheaper version, optimised for coding tasks without sacrificing much in performance. It’s perfect for developers needing a cost-effective yet potent model.

Both variants are available in ChatGPT and via OpenAI's API. O1-mini provides a balance between efficiency and power for developers who require strong reasoning capabilities but don’t need extensive world knowledge.

Performance Benchmarks:

The O1 model has demonstrated exceptional capabilities in problem-solving:

• Codeforces (Competitive Programming): 89th percentile
• AIME (USA Math Olympiad Qualifier): Top 500 students in the US
• GPQA (Physics, Biology, Chemistry): Exceeds human PhD-level accuracy
• International Olympiad in Informatics (IOI): 49th percentile globally
• Codeforces Elo Rating: 1807 (93rd percentile)

These results showcase O1's advanced reasoning abilities, enabling it to excel in complex tasks like developing sophisticated algorithms and conducting in-depth comparative analysis.

Challenges and Limitations:

Despite its strengths, O1 comes with higher operational costs—input costs are 3 times and output costs 4 times higher than GPT-4o on the API. It can also be slower, taking over ten seconds for complex queries. Moreover, O1 currently lacks features like web browsing and file analysis, and is prone to more frequent hallucinations and confident inaccuracies compared to its predecessors.

Access and Future Plans:

Currently, O1 models are available to ChatGPT Plus and Team users, with message limits of 30 per week for O1-preview and 50 for O1-mini. Enterprise and educational users will gain access next week, and developers with API usage tier 5 can start prototyping immediately. OpenAI plans to roll out O1-mini access to all free ChatGPT users in the future and is committed to refining the models and integrating additional features like browsing and file uploads.

Read more: OpenAI website | First impressions of OpenAI o1: An AI designed to overthink it

?? This Week's Spotlight: Apple Intelligence

Apple has introduced Apple Intelligence, a new personal intelligence system, alongside the iPhone 16 lineup. Here are the key details:

Apple Intelligence

Apple Intelligence is a new AI-powered system built into iOS 18, iPadOS 18, and macOS Sequoia. It aims to help users write, express themselves, and get things done more easily while protecting privacy.Key features include:

• ?? Apple Intelligence: Introducing new generative models and tools that integrate seamlessly with your iPhone for a more personalised experience.
• ?? Writing Tools: Allows rewriting texts, adjusting tones, proofreading, and even creating new emoji or images by typing descriptions.
• ??? Image and Video Search: Find specific moments or create movies by describing what you want to see, leveraging AI-powered searches.
• ?? Prioritising Communication: Summarises emails, notifications, and provides smart priority notifications to help focus on essential tasks.
• ?? Camera Control Feature: Instantly get detailed information about things seen through the camera, from restaurant details to event schedules.
• ?? Siri's Evolution: Enhanced language understanding, supporting voice stumbles, discreet typing requests, and step-by-step guidance on various tasks.
• ?? Private Cloud Compute: Ensures privacy while running advanced AI models, with independent verification to maintain data security.

Apple Intelligence will start rolling out in October 2024 for US English, with more languages coming later. However, in Australia it will roll out in December.

Read more: Apple Intelligence on iPhone in 5 minutes | Apple

?? Key Statistics

• 57% of enterprises are currently using Generative AI.
• 82% of organisations view Generative AI as a top priority for their AI strategy.
• 43% of AI leaders are deploying Generative AI at scale.
• 71% of companies using Generative AI report significant benefits.
• 65% of enterprises prefer cloud-native platforms for AI workloads.
• 88% of AI leaders use multiple cloud providers.

Read more: PR Newswire.

?? ChatGPT & OpenAI News

• ?? AI Adoption in Workplaces Surges: ChatGPT Reaches Over 1 Million Business Users: ChatGPT has crossed a significant milestone with over 1 million business users, demonstrating the rapid integration of AI in workplaces. Companies leverage ChatGPT for tasks like drafting emails, coding, and data analysis, enhancing productivity and decision-making processes. This growth reflects AI's expanding role in various business operations and its impact on efficiency. Read more: ZDNet
• ?? ChatGPT Outage Reported on September 11, 2024: ChatGPT experienced a global outage on September 11, 2024, causing widespread disruptions. Users reported being unable to access the platform, leading to significant community discussions. OpenAI acknowledged the issue and worked on a resolution to restore services. The cause of the outage and full details are yet to be disclosed, but users are advised to stay tuned for updates. Read more: DesignTAXI
• ?? OpenAI's New Sora Videos Showcase Stunning Human Realism: OpenAI has released new videos of Sora, an AI model that achieves remarkably lifelike human realism. These videos highlight advanced AI capabilities in replicating human expressions, gestures, and movements, pushing the boundaries of what's possible in synthetic media. Sora’s realism opens up new possibilities for applications in virtual avatars, simulations, and digital media, but it also raises ethical considerations around deepfakes and digital identity. Read more: Tom's Guide
• ?? OpenAI Invests in Drug Discovery Startup, Raises Nearly $30 Million: OpenAI has funded a drug discovery startup, raising almost $30 million to advance AI-driven solutions in the pharmaceutical sector. The investment aims to leverage AI's potential in speeding up drug discovery processes, identifying new therapeutic candidates, and optimising drug development. This move signifies a growing intersection of AI and biotechnology, where AI technologies are being increasingly applied to revolutionise healthcare and medicine. Read more: ReadWrite

? Claude Sonnet News

• ??? Claude AI Gets a Major Update: Code Highlighting and Editing: Claude AI has introduced a new update allowing users to highlight and edit code directly using text commands. This feature enhances the AI’s utility for developers by making code modification more intuitive and efficient, providing a more interactive experience when working with code snippets. Read more: Tom's Guide

?? Corporate AI Developments

• ?? Google's AI Note-Taking App 'NotebookLM' Adds Voice Explanations for Complex Topics: Google's AI-powered note-taking app, NotebookLM, now has a new feature that allows it to explain complex topics out loud. This addition aims to enhance user understanding by offering audible, AI-generated explanations for intricate subjects. The app continues to evolve, integrating advanced AI capabilities to provide users with a more interactive and educational experience. Read more: TechCrunch
• ?? Mistral Introduces 'PixTral': Its First Multimodal AI Model: Mistral has launched PixTral, its first multimodal AI model capable of processing both text and images. The model aims to integrate advanced visual and language understanding, enhancing applications in fields like content generation and interactive AI. This marks Mistral's entry into the competitive landscape of multimodal AI, positioning it alongside leading AI companies innovating in combining text and visual data processing. Read more: TechCrunch

?? AI Innovations

• ?? Replit: An Online Collaborative Coding Platform: Replit is a powerful browser-based platform for coding and collaboration, supporting multiple programming languages. It allows developers, educators, and teams to write, run, and share code seamlessly. With features like real-time collaboration, debugging tools, and deployment options, Replit is designed for both beginners and professionals looking to build and collaborate on software projects easily. Read more: Replit

?? Applications of AI

• ?? Connectly Leverages AI for Personalised Customer Messaging: Alibaba-backed Connectly is utilising AI to enhance customer communication by personalising messages across multiple platforms like WhatsApp and SMS. The company's AI-driven approach aims to improve customer engagement and retention by delivering tailored content, making interactions more relevant and impactful for businesses. Read more: TechCrunch
• ?? Bill Gates Uses AI for Meeting Summaries: Bill Gates frequently uses AI, particularly Microsoft's Copilot in Teams, to summarise meetings and extract key details. This feature allows him to efficiently review and interact with meeting content, saving time on note-taking and data entry. Gates highlights the benefits of AI in improving productivity, similar to how many workers use AI tools to save up to two hours daily on routine tasks. Read more: Business Insider
• ?? How to Use AI Tools Like ChatGPT, Claude to Save Time at Work: Workers across various fields, including law, marketing, and software development, are using AI tools like ChatGPT, Claude, and Jasper to save time on tasks such as summarising documents, drafting emails, conducting research, and improving code. A survey found that 75% of knowledge workers use AI at work, with 90% reporting time savings. Many see AI as a valuable tool to boost productivity, despite concerns about its limitations. Read more: Business Insider
• ?? How to Use ChatGPT to Automate Tasks: ChatGPT can automate various tasks, such as drafting emails, summarising content, managing schedules, generating creative content, and handling repetitive customer queries. By setting up the right prompts and workflows, users can integrate ChatGPT into their daily routines to boost productivity and reduce manual workload, making it a versatile tool for personal and professional use. Read more: Geeky Gadgets
• ?? Five ways professional writers are leveraging ChatGPT: OpenAI provides a detailed overview of how ChatGPT can elevate your writing. It covers various use cases, including brainstorming ideas, refining drafts, and even generating full-length articles. The guide is aimed at helping users leverage AI to enhance creativity, clarity, and efficiency in writing. Read more: OpenAI

AI for Accessibility and Inclusivity

• ?? The Power of AI in Empowering the Visually Impaired: Artificial Intelligence (AI) is transforming accessibility, particularly for the visually impaired. "Seeing AI," developed by Microsoft engineer Saqib Shaikh, uses AI to describe surroundings through a smartphone camera, enhancing independence for users with blindness or low vision. By leveraging computer vision and machine learning, Seeing AI reads text, identifies objects, and describes scenes, offering real-time, personalised audio feedback. This inclusive design not only empowers its users but also sets a new standard for future assistive technologies. Watch on YouTube

AI in Recruitment and Employment

• ?? AI Could Replace Traditional Job Interviews: AI is poised to revolutionise hiring by potentially eliminating traditional job interviews. Advanced AI can analyse candidates' skills and fit through simulations, automated assessments, and data analysis, reducing biases and inefficiencies in recruitment. While this may streamline the hiring process, it raises concerns about privacy and the human element in decision-making. Read more: Information Age

AI in Education

• ?? Integrating Generative AI into Australian Schools Becomes a National Priority: Australia is prioritising the integration of generative AI into its educational system. The initiative aims to enhance learning and teaching by leveraging AI tools to provide personalised education, support teachers, and prepare students for a future with AI-driven technologies. This move emphasises a balanced approach to adopting AI in schools, ensuring ethical use and addressing privacy concerns. Read more: Information Age

? Responsible AI

• ?? Albanese Government to Outlaw Doxxing in Australia: The Albanese government plans to criminalise doxxing—publishing someone's private information online without consent—under new legislation. This move aims to protect individuals' privacy and curb online harassment. The proposed law will impose penalties on those found guilty of doxxing, reflecting the government’s commitment to tackling cyber abuse and enhancing digital safety. Read more: ABC News
• ??? Senate Leaders Urge FTC to Investigate AI Content Summaries for Anti-Competitive Practices: U.S. Senate leaders have asked the Federal Trade Commission (FTC) to investigate whether AI-generated content summaries could harm competition. The concern is that large companies might monopolise AI tools to dominate markets, potentially disadvantaging smaller content creators and publishers. The request reflects broader worries about the impact of AI on digital content markets and fair competition. Read more: TechCrunch
• ?? ChatBLACKGPT Founder Erin Reddick on AI for Black Communities: Erin Reddick, founder of ChatBLACKGPT, discusses her vision to create AI tools tailored to the unique needs and experiences of Black communities. In an interview, she highlights the importance of representation in AI development and the potential for AI to empower marginalised voices through culturally sensitive responses and applications. Read more: Mashable
• ?? Meta Faces Scrutiny Over Using Australian Content to Train AI: Meta is under fire for using Australian social media posts to train its AI models without user consent. Concerns have been raised about privacy and ethical practices, as this data collection may infringe on user rights. Australian authorities are questioning Meta's methods and seeking clarification on how data is being used to develop AI technologies. Read more: Information Age

?? AI Learning Hub

• ?? Understanding AI Prompt Engineering: Crafting Effective Instructions for AI Models: AI prompt engineering involves designing and refining text inputs to guide AI models like GPT or Claude in generating accurate, relevant responses. It combines natural language processing, system design, and iteration to optimise AI outputs. Prompt engineers develop prompts, test them extensively, and integrate them into broader systems, considering factors like error handling and feedback loops. As AI advances, prompt engineering is expected to evolve, potentially becoming more collaborative and interdisciplinary. Watch on YouTube
• ?? Understanding Retrieval-Augmented Generation (RAG) for Businesses: RAG integrates large language models (LLMs) with business knowledge by combining external data retrieval with generative AI. This approach enhances AI's ability to provide relevant, accurate responses based on specific business contexts. RAG can improve customer service, data analysis, and content creation by leveraging up-to-date, domain-specific information alongside AI's natural language generation capabilities. Learn more: ZDNet
• ?? The Future Impact of Artificial Intelligence According to Elon Musk: Elon Musk envisions AI creating an "age of abundance" by reducing production costs and automating tasks, leading to near-zero costs for goods and services. Key highlights include the rise of autonomous vehicles, humanoid robots like Tesla's Optimus, and the potential "crisis of meaning" if AI performs all human jobs better. Tesla's advanced AI infrastructure, with its Dojo supercomputer and custom AI chips, is set to drive this transformation, making Musk’s vision both revolutionary and ethically complex. Watch on YouTube

??? We'd Love to Hear from You!

Did you find this edition insightful? Like, share your thoughts, and ask us anything.

Stay curious, and until next time, enjoy exploring the vast possibilities AI offers.

???? Subscribe on LinkedIn The ChatGPT Observer Newsletter

As we explore the captivating world of AI technologies, it's important to note that the content of this article is a product of human effort, enhanced by AI tools and additional resources. We have made sure to properly attribute and link all referenced articles to their original sources. For the visuals:

• Images sourced from external sites are credited within the article.
• Other images were generated using DALL-E, MidJourney and IdeoGram with enhancements made in Canva. These AI-generated and enhanced images are designed to visually represent each section's key themes.

What do you think about the AI-generated images in this edition? We'd love to hear your thoughts on how they enhance your reading experience!