On the Cusp of Superintelligence chronicles the most transformative year in artificial intelligence history. In 2025, every assumption about how AGI would emerge was shattered. What began with OpenAI's o3 model scoring 87.5% on the ARC-AGI benchmark, a test designed to measure genuine reasoning that the average human scores 85% on, became a cascade of breakthroughs that fundamentally reshaped the race toward superintelligence.

This book reveals how five competing visions of artificial general intelligence are converging simultaneously. OpenAI bet everything on scaling, pushing from GPT-3's 175 billion parameters to o3's reasoning breakthrough and GPT-5's hybrid architecture. DeepMind, led by neuroscientist Demis Hassabis, pursued a portfolio strategy combining hierarchical reasoning, self-improving systems, and scientific discovery. Anthropic built safety directly into its architecture through Constitutional AI, proving that capability and alignment need not be mutually exclusive. Meta championed world models and embodied intelligence, arguing that language models without physical grounding are like philosophers in Plato's cave. And xAI's Grok series, trained on the massive Colossus supercluster with 200,000 GPUs, pushed the boundaries of truth-seeking through real-time information integration.

Then came the shock no one expected. In January 2025, China's DeepSeek-R1 matched Western AI performance using a fraction of the computational resources, creating what researchers called an "AI Sputnik moment." Through pure reinforcement learning and Mixture-of-Experts architectures, Chinese engineers proved that the path to AGI might require clever engineering rather than brute-force scale.

As these architectural philosophies race forward, On the Cusp of Superintelligence confronts the deeper crisis: the alignment problem. How do we ensure superintelligent systems pursue human goals rather than developing their own? Models classified at Anthropic's Level 3 risk tier already show autonomous tendencies their creators did not anticipate. The question is no longer whether AGI will arrive, but whether we will be ready when it does.

This book is for readers who want to understand:

• How OpenAI, DeepMind, Anthropic, Meta, and xAI are each pursuing different paths to artificial general intelligence
• Why China's DeepSeek breakthrough changed the global AI race overnight
• What the ARC-AGI benchmark reveals about machine reasoning versus human intelligence
• How the alignment problem evolved from theoretical concern to immediate crisis
• What hybrid architectures, recursive self-improvement, and world models mean for the future of superintelligence

Written by H. Peter Alesso, author of Self-Improving AI, AI Operating Systems, and Absolute Zero Reasoner, with publications from Wiley-IEEE Press and extensive expertise in artificial intelligence research. This is the essential guide for anyone who wants to understand where the AGI race stands and what comes next.

Start reading today and discover why 2025 was the year everything changed.