The fervor surrounding artificial general intelligence (AGI) has significantly waned in recent months. Just a few months ago, top tech executives were touting the imminent arrival of AGI, with some predicting it could happen as soon as 2025 or 2026. However, many are now downplaying their claims and instead focusing on more practical applications of AI.
The shift away from AGI rhetoric is attributed to growing concerns that progress in AI development may not be as rapid as initially thought. The rollout of OpenAI’s GPT-5 model in early August failed to deliver the breakthroughs expected, with many experts describing it as incremental improvements without significant advancements.
Many industry leaders have acknowledged that the term “AGI” is overhyped and lacks a clear definition. AI researcher Ben Goertzel defined AGI as an AI system capable of learning to reliably perform any cognitive task that a competent human can, but even this definition has its problems.
The emphasis on AGI was initially driven by OpenAI’s mission to develop the technology for the benefit of all. However, with growing concerns about the actual progress being made, many are now focusing on more concrete applications where AI advancements can benefit society.
While some still raise urgent risks associated with developing superhuman AI systems, others argue that the retreat from AGI rhetoric is a necessary move to avoid regulatory scrutiny and ensure the development of practical AI solutions. The industry’s focus has shifted towards domain-specific “superintelligences” rather than a monolithic AGI fantasy.
The hype around AGI may be fading, but with so much at stake – including jobs, security, and safety – the real questions about where this race leads are only just beginning.
Source: https://fortune.com/2025/08/25/tech-agi-hype-vibe-shift-superpowered-ai