Current Progress and Challenges in the Race Towards AGI and Superintelligence

The article discusses the ongoing global race, primarily led by US and Chinese tech giants, to develop Artificial General Intelligence (AGI) and superintelligence. AGI is defined as AI capable of performing any human job autonomously, while superintelligence refers to AI far exceeding human cognitive abilities. OpenAI’s CEO Sam Altman described their latest GPT-5 upgrade as a significant progress toward AGI but noted it still lacks key capabilities such as continuous learning after deployment. Other companies including Meta, Google, Anthropic, and Chinese firms like DeepSeek are making advances, but experts warn that the race is currently driven more by competitive hype than scientific breakthroughs. Analysts highlight the absence of a theoretical model explaining how current generative AI could truly achieve AGI, comparing the effort to building Apollo rockets without understanding gravity. Some experts believe a narrower and more practical definition of AGI could be reached within five to ten years, but the bar for what constitutes AGI continues to evolve. Despite this, present AI systems already generate substantial revenues and improvements in capabilities such as reasoning and multimodal understanding. The article also stresses concerns about managing reliability, transparency, and ethical use instead of chasing uncertain superintelligence. The global contest includes heavy investments — nearly $400 billion by US tech firms in 2025 alone — with China also playing a major role in AI development and adoption worldwide. The pursuit reflects both technological ambition and geopolitical positioning rather than settled scientific certainty.