OpenAI, Anthropic reveal how users use AI

INSIGHTS FROM OPENAI & ANTHROPIC

Summary: OpenAI and Anthropic have released fresh data revealing how millions engage with their AI platforms. The findings highlight demographic shifts, regional disparities, and a growing divide between personal and professional AI use.

• Claude users predominantly utilize AI for coding tasks, whereas ChatGPT is favored for writing assistance and decision-making support.
• Personal interactions with ChatGPT surged from 53% in mid-2024 to 73% in early 2025, indicating a rapid increase in non-work-related conversations.
• Adoption rates in low- and middle-income countries for ChatGPT are expanding at four times the pace compared to wealthier regions, where Claude remains more popular.
• Both platforms show a trend toward users delegating more tasks to AI, with a notable rise in information retrieval and search-oriented queries rather than content generation.

Significance: These insights underscore the accelerating global embrace of AI, revealing nuanced differences in how various populations integrate AI into their lives. The surge in personal use suggests AI’s transformative impact extends well beyond professional environments.

ENHANCING SOFTWARE DEVELOPMENT WITH AUGMENT CODE

Overview: Augment Code offers a sophisticated AI-powered coding assistant designed to support developers by providing deep contextual understanding of complex codebases and automating routine tasks.

• Effortlessly index and explore millions of lines of code.
• Receive immediate, precise answers about any segment of your code.
• Streamline workflows by automating processes across your entire development environment.

This tool is tailored to enhance collaboration and efficiency for teams managing large-scale software projects.

OPENAI’S GPT-5 CODEX: A NEW ERA IN AUTONOMOUS CODING

Key Update: OpenAI has unveiled GPT-5 Codex, an advanced coding model that dynamically allocates computational resources based on task complexity, enabling it to handle everything from quick fixes to multi-hour problem-solving sessions.

• Outperforms the standard GPT-5 on the SWE-bench Verified benchmark, especially excelling in code refactoring with a 51.3% success rate compared to 33.9%.
• Reduces token consumption by 94% for straightforward tasks while doubling reasoning time for intricate challenges, supporting autonomous operations lasting over seven hours.
• Incorporates comprehensive code review features that analyze entire codebases, run tests, and verify dependencies to detect critical bugs.
• Includes updated command-line tools, IDE plugins for VS Code and Cursor, and seamless transitions between local and cloud development environments.

Why This Matters: The evolution of agentic coding assistants like GPT-5 Codex is revolutionizing software development workflows. Despite early dominance by Anthropic’s Claude Code, OpenAI’s advancements position it as a formidable competitor with a broad user base.

BOOST YOUR CREATIVE THINKING WITH GPT-5

Overview: Discover how to break free from repetitive thought patterns and generate innovative solutions using GPT-5’s structured inquiry method. This approach helps uncover hidden challenges and stimulates fresh perspectives for any business dilemma.

How to Apply:

1. Activate GPT-5 in Auto mode and prompt it with: “I’m trying to [describe your goal or problem], but I keep circling the same ideas. Ask me enough questions to help find a new approach.”
2. Provide background on your current strategies so GPT-5 can tailor its questions to push beyond existing solutions.
3. Respond to GPT-5’s inquiries across areas like user insights, product experience, engagement, and analytics to reveal novel angles.
4. Evaluate the new ideas generated and iterate on the most promising ones.

Pro Tip: This technique is versatile and effective for any challenge, not just customer churn. The power lies in GPT-5’s persistent questioning that guides you toward breakthrough thinking.

SPONSORED INSIGHT: FIDDLER’S AGENTIC OBSERVABILITY GUIDE

Fiddler’s comprehensive guide explores how to monitor and manage complex multi-agent AI systems in production environments.

• Identify issues at various levels within agentic hierarchies.
• Visualize both aggregated metrics and detailed insights.
• Understand the decision-making pathways and logic chains of AI agents.

REVE: REVOLUTIONIZING IMAGE CREATION AND EDITING

Overview: Reve has launched an innovative image platform that merges AI-driven image generation, natural language editing, and intuitive drag-and-drop controls into a single, free-to-use interface.

• Utilizes a “layout representation” system that translates images into code-like structures, enabling precise modifications while maintaining the integrity of the original image.
• Features a drag-and-drop editor for detailed adjustments to both uploaded and AI-generated images.
• Includes a chat interface allowing users to create, blend, and edit images through natural language commands, with integrated web search for inspiration.
• Offers beta API access for developers to embed Reve’s image creation and editing capabilities into external applications and workflows.

Why It’s Important: Following recent breakthroughs like Google’s Nano Banana and ByteDance’s Seedream 4.0, Reve’s advanced editing tools mark the next frontier in image AI, unlocking new creative possibilities beyond mere generation.

QUICK UPDATES

• OpenAI’s latest model enhances agentic coding efficiency.
• Google releases VaultGemma, an AI model prioritizing user data privacy through differential privacy techniques.
• A new free AI image generator debuts with sophisticated editing features.
• H Company introduces Holo 1.5, a suite of open-weight foundation models excelling in agentic benchmarks.

INDUSTRY NEWS

Upcoming Event: Join the AWS Symposium in Ottawa on October 1 to connect with public sector innovators, explore emerging technologies, and influence the future of AI in Canada.

Google: Launches VaultGemma, the largest publicly accessible AI model trained with differential privacy to safeguard user information.

OpenAI: Expands its robotics team with experts in humanoid systems, signaling a strategic move into physical AI applications.

H Company: Releases Holo 1.5, a new family of open-weight models achieving state-of-the-art results on agentic AI benchmarks.

Microsoft: Integrates Copilot Chat and AI agents across its Office 365 suite, embedding AI tools directly into applications like Word and Excel for seamless user access.

OpenAI Chairman Bret Taylor: Warns of an AI investment bubble with potential financial losses but emphasizes the technology’s capacity to generate substantial economic value.

COMMUNITY SPOTLIGHT

Reader Workflow: Mark D. from the UK

“To make my two sports-loving sons more engaged with their school reports, I transformed the usual dry PDFs into dynamic soccer match analyses using ChatGPT. The reports included commentary on their ‘stats,’ highlights, and ‘training areas.’ I then used ElevenLabs to give the scripts the voice of an enthusiastic sports commentator.

The outcome? School feedback delivered like a Match of the Day highlight reel. Strengths became ‘goals scored,’ and areas for improvement turned into ‘training drills.’ For the first time, my boys asked to listen to their reports multiple times.”

How are you leveraging AI? Share your story with us!

RESOURCES & UPCOMING EVENTS

• Catch up on our previous AI newsletter editions.
• Explore the latest in technology and robotics newsletters.
• Discover today’s curated AI tool guide.
• Reserve your spot for the next workshop, happening Friday at 4 PM EST.