OpenAI launches image generation as Mahmoud Malaeb unveils GEPA

Key Takeaways

• OpenAI's ChatGPT now generates and edits images from text descriptions, allowing for detailed refinements and the use of multiple reference images.
• Robert Bunn developed Delivery Date AI, an FDA-approved technology that analyzes ultrasound images to predict baby due dates and provide early warnings for complications.
• Residents near Microsoft's new AI data center in Mount Pleasant, Wisconsin, are experiencing persistent humming noise and Microsoft is investigating.
• The D.C. region launched Talent Capital AI, a jobs board with over 83,000 users and 65,000 jobs, offering free virtual training in AI skills.
• The difficult job market for the Class of 2026 is primarily attributed to financial factors and interest rate changes, not AI.
• Governments need to publish information in AI-readable formats to ensure accuracy from chatbots providing public services.
• Therapists should inquire about patients' use of AI chatbots like ChatGPT for emotional support to better understand their needs and provide guidance.
• A survey indicates that 92 percent of clinicians prioritize human listening over AI for diagnosis, despite challenges like time pressure.
• Argenta CEO Mahmoud Malaeb introduced GEPA, a new iterative framework for evaluating AI agents that learns from mistakes to improve reliability.

ChatGPT Now Creates Images From Text

OpenAI's ChatGPT can now generate images based on text descriptions, making it easier for users to create visual content quickly. The tool allows for detailed adjustments to composition, size, and style by refining text prompts. Users can also edit existing images by providing specific instructions for changes. ChatGPT supports using multiple uploaded images to guide the creation or editing process. This feature aims to streamline visual content creation for various purposes.

ChatGPT's Image Generation Tool Explained

OpenAI's ChatGPT now creates images from text, simplifying visual content creation. Users can quickly generate and refine images by providing clear, detailed text prompts. The system allows for specific edits to composition, style, and elements within the image. It also supports using multiple images as references for generation or modification. This advanced capability aims to enhance visual communication and asset creation.

Colorado Father's AI Predicts Baby's Due Date

A Colorado father developed AI technology called Delivery Date AI to predict a baby's arrival date more accurately than traditional methods. The system analyzes ultrasound images to estimate due dates, potentially offering earlier warnings for pregnancy complications. This technology was inspired by the father's personal loss due to premature births. Delivery Date AI has received FDA approval and is now available for use by doctors in the United States.

AI Due Date Predictor Developed After Personal Tragedy

Robert Bunn, a Colorado father, created Delivery Date AI after experiencing miscarriages, aiming to improve pregnancy care. This AI technology analyzes ultrasound images to predict a baby's due date with high accuracy, potentially alerting doctors to issues sooner. Unlike traditional methods, it looks for patterns that indicate potential problems. The technology, inspired by Bunn's personal loss, has gained FDA approval and is being used to help prevent surprise premature births.

Therapists Should Ask Patients About AI Use

Experts suggest therapists should ask patients if they use AI chatbots like ChatGPT for emotional support. This helps understand why patients seek AI companionship and what they gain from it. Discussing AI use can reveal underlying issues, such as avoiding difficult conversations or coping with anxiety and depression. Therapists can use this information to provide better support and guide patients on AI's limitations. Approaching the topic with curiosity, not judgment, is key.

Microsoft AI Data Center Causes Persistent Hum

Residents near Microsoft's new AI data center in Mount Pleasant, Wisconsin, are complaining about a persistent humming noise. The facility, which will train AI models, is set to become operational later this year. Neighbors expressed frustration during a Q&A session with Microsoft representatives, citing disturbances like headaches and difficulty sleeping. Microsoft acknowledged the noise concerns and stated they are investigating the root cause.

Class of 2026 Job Struggles Not Caused by AI

The Class of 2026 is facing a difficult job market, but experts say AI is not the primary reason. The main drivers are financial factors, including low interest rates that encouraged spending and subsequent steep rate hikes. These economic shifts have led to an entry-level hiring freeze, similar to past economic downturns. While AI may transform jobs in the future, it is not currently responsible for the hiring challenges faced by recent graduates.

DC Region Launches Jobs Board for Future Employment

The D.C. region has created a new jobs board called Talent Capital AI to help residents prepare for a changing job market. The platform assists displaced workers and aims to connect people with training opportunities, especially in areas like project management, leadership, and AI skills. Currently, it has over 83,000 active users and lists more than 65,000 jobs, with healthcare and engineering leading the sectors. The initiative also offers free virtual training in foundational AI skills.

Government Information Needs to Be AI Readable

Artificial intelligence chatbots are becoming a primary way people access government services, but they often provide incorrect information. This happens because government guidance is frequently outdated, fragmented, or not easily readable by machines. Experts urge governments to publish information in formats that both humans and AI can understand. Making information machine-readable is crucial to ensure AI assistants provide accurate guidance and prevent errors for citizens seeking help.

Human Listening Skills Trump AI in Medicine

A survey reveals that 92 percent of clinicians prioritize listening over AI for diagnosis, with many identifying critical conditions solely through sound. Despite AI advancements, human listening remains the essential diagnostic tool. However, clinicians report that time pressure and high patient volumes hinder their ability to listen effectively. Experts emphasize that listening is a crucial clinical skill that builds trust and centers the patient, and medical education needs to reinforce its importance.

New Framework Improves AI Agent Evaluation

Argenta CEO Mahmoud Malaeb introduced GEPA, a new framework for evaluating AI agents more accurately. Traditional AI judges often fail to understand complex policies, leading to incorrect compliance assessments. GEPA uses an iterative process where the AI learns from its mistakes, improving its evaluation rubric over time. This data flywheel approach helps build more reliable AI evaluators, especially for customer-facing roles. The system was demonstrated using the Tau-2 benchmark dataset.

Sources

• OpenAI's ChatGPT Now Generates Images
• ChatGPT's New Image Generation
• Colorado father's loss inspires AI technology that could change pregnancy care
• A Colorado father's loss inspires AI technology that could change pregnancy care
• 'How Are You Using AI?' Therapists Should Ask You That Question, Experts Argue
• Microsoft holds Q&A about Mount Pleasant AI data center, doesn't offer solution to persistent hum
• The Class of 2026 is struggling to find jobs—and it’s not because of AI
• From DOGE to AI: DC region’s newly created jobs board seen as tool for future employment
• AI Is the New Front Door to Government. The Bots Need Help.
• Why clinical listening skills outpace artificial intelligence
• LLM Evaluators: Beyond Naive Judgments