# OpenAI talks GPT-6 Astra and Millenium Prize, researchers create WeWorm exploit & IBM’s US Open app

## Executive summary

This episode reviews rapid advancements in AI, covering OpenAI's GPT-6 Astra model, its application in solving the Millennium Prize problem (Navier-Stokes), and the use of AI for complex tasks like 3D rendering and code generation. The discussion also details IBM's use of AI for personalized sports analytics at the US Open, including biomechanical limb tracking and real-time match forecasting. Finally, the segment warns about the increasing accessibility of advanced cyber threats, exemplified by the WeWorm exploit, which uses AI to discover zero-click vulnerabilities in messaging apps.

## Key takeaways

- GPT-6 Astra and AI's Role in Solving Math Problems: OpenAI introduced GPT-6 Astra, a model capable of state-of-the-art performance across computer use, software engineering, and cybersecurity. Astra reportedly solved the Navier-Stokes equation, a Millennium Prize problem, using 10,000 AI agents and requiring massive computational resources (estimated at $15 million worth of compute for a $1 million prize).
- IBM's AI for US Open Analytics: IBM enhanced the US Open experience by implementing AI features, including pre-match likelihood forecasting (using classical ML models like boosted trees and logistic regression), real-time match probability tracking, and 'Serve Quality' metrics derived from tracking 21 points across a player's body at 50 frames per second.
- WeWorm: AI-Assisted Exploits: A security firm used AI to discover a zero-click exploit (WeWorm) in WeChat's VoIP stack. This worm can autonomously spread and infect contacts simply by receiving a call, highlighting the growing threat of AI-assisted offensive security.

## Technical details

- Agent Models and Computational Overhead: The discussion questioned the efficiency of large agent models, noting that while they can perform complex tasks (e.g., ordering DoorDash), the computational cost (tokens used) and the method of problem-solving must be critically evaluated for real-world enterprise use cases.
- US Open Predictive Modeling: Match forecasting utilizes a combination of classical machine learning models (Boosted Trees and Logistic Regression) alongside media sentiment analysis and head-to-head statistics to predict match outcomes. Live probability is updated in real-time based on point-by-point outcomes.
- Biomechanical Tracking and Serve Quality: The 'Serve Quality' metric tracks 21 points across a player's body at 50 frames per second. This process generates massive datasets (e.g., 1.2 billion data points over a tournament) used to analyze the biomechanics of the serve across distinct phases (start, loading, cocking, acceleration, contact, deceleration).
- Cybersecurity Vulnerabilities: The WeWorm exploit was found to be a memory corruption issue in the WeChat VoIP stack. The vulnerability was discovered using AI and represents an 'unconventional attack surface,' suggesting that AI is exposing weaknesses in complex, dynamic software systems.

## Practical implications

- AI-driven agent models are moving beyond simple Q&A to performing complex, multi-step tasks (e.g., generating executable CAD code, analyzing biomechanics).
- The integration of classical ML models (like logistic regression) with advanced AI capabilities is transforming consumer-facing data products (e.g., sports analytics).
- Build engineers must account for the increasing complexity of 'unconventional attack surfaces' in modern, dynamic, and emergent software systems.
- The economic cost of AI computation is becoming a critical factor in determining the feasibility and utility of advanced AI solutions.

## Topics

Generative AI, Cybersecurity, Machine Learning, Sports Technology, Computational Mathematics, GPT-6 Astra, Navier-Stokes Equation, WeChat, Boosted Trees / Logistic Regression

Source: https://www.youtube.com/watch?v=XPReiOKCzFI