OpenAI is significantly enhancing its ChatGPT platform with several new features designed to improve user experience and organization. The company introduced ChatGPT Projects, creating dedicated spaces for users to organize chats, files, and instructions, preventing scattered information. Additionally, Custom Instructions allow users to set a default working style, while the Memory feature enables ChatGPT to recall specific details, making interactions more personalized and efficient.
Further expanding personalization, OpenAI now allows users to build their own custom GPTs, which are tailored ChatGPT assistants for specific tasks, configurable with instructions, knowledge files, and tools like web search. ChatGPT is also being highlighted as a powerful writing assistant for drafting and revising content. In a practical application, Derek Wallbank successfully used ChatGPT as an AI coach and nutritionist to prepare for the Paris Marathon on April 12, 2026, leveraging his fitness data for a personalized training plan. OpenAI also introduced a new mid-tier subscription plan for its coding agent, Codex, offering five times more usage than the Plus subscription.
Meanwhile, Meta has launched Muse Spark, its most powerful proprietary AI model from Meta Superintelligence Labs, aiming to compete directly with Google and OpenAI. Muse Spark is designed for faster reasoning and more personalized, context-aware answers, capable of processing text and visual inputs. Currently available in the Meta AI app and website in the US, Meta plans to integrate it into other platforms like Facebook and WhatsApp, and even AI glasses, while exploring revenue streams through API sales.
Beyond these major players, AI is finding diverse applications and raising important discussions. Autodesk launched a read-only MCP server, allowing external AI agents secure access to documentation for over 110 products, enhancing accuracy. In healthcare, Deepesh Randeri, CISO at Akron's Children's Hospital, stresses the critical need for awareness and governance in AI security, including due diligence on vendors and monitoring for bias. AI is also transforming golf, from booking tee times to course maintenance.
The debate around AI regulation is also intensifying, with tech executives from companies like Meta, Anthropic, and OpenAI actively influencing discussions in Congress. OpenAI's Sam Altman has proposed "people-first policy ideas" for regulation. While some argue against industry-authored regulations due to concerns about shareholder capitalism, others believe policymakers require industry input to develop effective frameworks, despite ongoing trust issues.
Key Takeaways
- OpenAI introduced ChatGPT Projects to help users organize chats, files, and instructions in dedicated spaces.
- ChatGPT now includes Custom Instructions for setting default working styles and a Memory feature to recall specific user details.
- Users can build custom GPTs, which are personalized ChatGPT assistants tailored for specific tasks with custom instructions and knowledge.
- OpenAI launched a new mid-tier subscription for its Codex coding agent, providing five times more usage than the Plus plan.
- Meta released Muse Spark, its most powerful proprietary AI model, to compete with Google and OpenAI, offering faster reasoning and personalized answers.
- Muse Spark, developed by Meta Superintelligence Labs, processes text and visual inputs and is available in the Meta AI app, with plans for broader integration.
- Autodesk launched a read-only MCP server, enabling external AI agents to securely access product documentation for over 110 products.
- Deepesh Randeri, CISO at Akron's Children's Hospital, emphasized the importance of awareness, governance, and monitoring for bias in AI security within healthcare.
- AI is transforming the golf industry, from managing tee time reservations to optimizing course maintenance and detecting turf disease.
- Tech executives from Meta, Anthropic, and OpenAI, including Sam Altman, are actively influencing congressional discussions on AI regulation.
OpenAI adds Projects to ChatGPT for better organization
OpenAI has introduced ChatGPT Projects, a new feature that creates dedicated spaces to organize chats, files, and instructions. This helps users keep related information together for ongoing AI-assisted work. Projects aim to solve the problem of scattered information across different conversations. Users can create a project, name it, and add files, custom instructions, or existing chats. For some plans, collaborators can be invited to work together on projects.
ChatGPT now remembers user preferences with new features
OpenAI is improving ChatGPT with Custom Instructions and Memory features to make interactions more personalized and efficient. Custom Instructions let users set a default working style, like preferred tone or output format, which stays with them across chats. The Memory feature allows ChatGPT to recall specific details users ask it to remember, reducing the need to repeat information. These tools help turn ChatGPT into a more adaptable partner for users.
OpenAI allows users to create custom AI assistants
OpenAI is enabling users to build their own custom GPTs, which are personalized ChatGPT assistants designed for specific tasks. These custom assistants can be tailored with specific instructions, knowledge files, and tools like web search or data analysis. Users can create them through a conversational interface or a detailed configuration tab. OpenAI provides examples like a writing coach and a data analyst, and users can test and refine their creations before sharing them.
Meta launches Muse Spark AI for enhanced product features
Meta has released Muse Spark, the first AI model from its Meta Superintelligence Labs, to improve AI features across its products. This model is designed to be faster and better at reasoning, aiming to provide more personalized and context-aware answers. Muse Spark can process text and visual inputs, analyze photos and social posts, and generate richer replies. It is currently available in the Meta AI app and website in the US, with plans to expand to other platforms like Facebook and WhatsApp.
Meta's Muse Spark AI challenges Google and OpenAI
Meta has launched Muse Spark, its most powerful proprietary AI model, to compete with Google and OpenAI. This model is integrated into Meta's AI app and will be added to other products like AI glasses. Meta plans to invest heavily in AI and explore revenue streams through API sales and paid access to advanced features. Muse Spark aims to overcome limitations of previous models like Llama and represents a significant step in Meta's AI development strategy.
ChatGPT becomes a powerful AI writing assistant
OpenAI is highlighting ChatGPT's capabilities as a professional writing assistant that can help with drafting, revising, and adapting content for different audiences. The process involves planning, drafting with AI, revising for clarity and tone, and packaging the final content. Users should provide clear context and constraints, and always review the AI's output. Specific prompts can be used for tasks like drafting emails or creating internal updates.
AI coach helps train for Paris Marathon
Derek Wallbank used ChatGPT as an expert running coach and nutritionist to prepare for the Paris Marathon on April 12, 2026. He provided the AI with years of his fitness data, including Strava logs and diet details, with the goal of finishing injury-free. The AI created a personalized training plan that included running, gym workouts, and nutrition advice. Wallbank logged his progress and provided feedback, allowing the AI to adjust the plan over the six months leading up to the race.
OpenAI launches new subscription for Codex users
OpenAI has introduced a new mid-tier subscription plan for its coding agent, Codex, due to high demand. This new plan offers five times more Codex usage than the Plus subscription, making it suitable for longer and more intensive coding sessions. OpenAI also rebalanced the usage limits in its existing Plus plan to better support daily use. The company previously introduced Codex in May 2025, and its weekly user base has grown significantly.
Autodesk server gives AI access to product documentation
Autodesk has launched a read-only MCP server that allows external AI agents to securely access its product help documentation. This service enables AI assistants to provide more accurate and reliable responses based on verified information. The server covers over 110 Autodesk products, helping users with onboarding, exploring features, and troubleshooting. This move integrates domain-specific knowledge with AI, improving the accuracy of guidance for industrial and enterprise applications.
AI security needs awareness and governance in healthcare
Deepesh Randeri, CISO at Akron's Children's Hospital, emphasizes the importance of awareness and governance for AI security in healthcare. He explains that all new technologies, including AI, must go through a structured process involving committees to ensure standard of care, cost optimization, ROI, and security. Randeri highlights the need for due diligence on vendors and systems, and stresses the importance of monitoring AI's back-end performance for bias and adherence to implemented guardrails.
AI leaders influence regulation debate
As Congress considers AI regulation, tech executives from companies like Meta, Anthropic, and OpenAI are actively influencing the discussion. OpenAI's Sam Altman has proposed 'people-first policy ideas' to guide regulation. Some argue that AI companies should be regulated rather than authoring regulations, citing shareholder capitalism. Others believe policymakers lack the necessary AI expertise and need industry input to create effective regulations, despite concerns about industry influence and trust.
AI is transforming golf for players and courses
Artificial intelligence is beginning to transform the game of golf, impacting everything from booking tee times to maintaining courses. AI-powered virtual assistants will soon handle tee time reservations, considering player data and preferences. On the course, AI is being used to manage fairways and greens, detect turf disease, and optimize watering. This technological integration promises to accelerate in the coming years, changing how golfers interact with the game and how courses are managed.
Sources
- OpenAI's ChatGPT Projects Organize Your AI Work
- ChatGPT Gets a Memory Upgrade
- OpenAI Lets Users Build Custom GPTs
- Meta Introduces Muse Spark to Strengthen AI Across Its Products
- Meta Unveils Muse Spark AI, Aims to Challenge Google and OpenAI
- ChatGPT: Your AI Writing Assistant
- I Trained for the Paris Marathon Using ChatGPT
- Codex Demand Sparks New OpenAI Mid-Tier Subscription for Power Users
- Autodesk Launches Product Help MCP Server: Enabling Trusted Documentation Access for AI Assistants
- AI security starts with awareness and governance
- Debatable: AI titans influencing regulation
- How AI Is Changing Golf for Players and on the Course
Comments
Please log in to post a comment.