🗞 News of the day

🤖 AI Reprograms Brain Cancer Cells

USC researchers are leveraging AI to develop a new immunotherapy approach for treating glioblastoma.

• AI-Driven Gene Identification: Researchers at USC Norris Comprehensive Cancer Center have developed a machine learning system that identifies genes capable of converting glioblastoma cells into dendritic cells (DCs). This innovative approach, detailed in the Cancer Immunology Research journal, allows for the analysis of tens of thousands of genes and their interactions, potentially revolutionizing glioblastoma treatment.

• Challenges of Glioblastoma: Glioblastoma is the most common and deadliest brain cancer in adults, with a survival rate of less than 10% after five years post-diagnosis. Current immunotherapy methods have limited success due to the blood-brain barrier that shields tumors and complicates immune responses. The new AI-based approach aims to overcome these obstacles by reprogramming cancer cells into DCs, which can activate the immune system more effectively.

• Promising Results and Future Directions: In mouse models, this AI-driven method increased survival rates by up to 75%. When combined with immune checkpoint therapy or classic DC vaccines, survival rates improved significantly. The next steps involve fine-tuning these genes for human application and conducting safety tests in animal models, with the goal of eventually expanding the research to other types of cancer and moving towards clinical trials.

This groundbreaking AI research offers hope for more effective glioblastoma treatments, though further development and testing are necessary before it can be used in clinical settings.

🔖 Today's Cool Story

🚨 OpenAI Begins Hyper-realistic Voice Rollout

OpenAI's new advanced voice mode, powered by the GPT-4o model, is set to revolutionize AI-human interaction with hyper-realistic audio responses.

• Launch and Features: The advanced voice mode will be available to ChatGPT Plus subscribers starting July 30, 2024, offering real-time conversations with emotional tone detection. The feature includes four preset voices and allows users to interrupt and engage naturally with the AI.

• Safety and Rollout: OpenAI has implemented robust safety measures, including testing with over 100 external red teamers in 45 languages, and filters to block inappropriate content. The rollout will be gradual, aiming for full access by the end of fall 2024.

• Background and Future Plans: The initial demo faced controversy due to a voice resembling Scarlett Johansson, leading to its removal. OpenAI plans to expand the capabilities of the advanced voice mode to include video and screen sharing in the future. A detailed safety report on the feature's testing will be released in early August.

🚀 Try New Human Resource Prompts

🎨Image Prompts

Prompt #1: Living room by the snow mountain

Prompt #2: Wooden cabin in nature

Prompt #3: Bright colors indoor decoration

See all Midjourney prompts →