Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

Rather than reaching philosophical benchmarks, general intelligence can be practically defined as models matching average B-player corporate employee performance across routine tasks. Under this pragmatic workplace standard, basic general intelligence is already present across many business functions.

Related Insights

Even a specialized task like coding involves a wide range of human-like interaction: brainstorming, searching, and more. This "AGI-completeness" means a powerful general model with a good "bedside manner" can outperform a narrowly specialized one, complicating the strategy for vertical AI apps.

A consortium including leaders from Google and DeepMind has defined AGI as matching the cognitive versatility of a "well-educated adult" across 10 domains. This new framework moves beyond abstract debate, showing a concrete 30-point leap in AGI score from GPT-4 (27%) to a projected GPT-5 (57%).

Dan Shipper proposes a practical, economic definition for AGI that sidesteps philosophical debates. We will have AGI when AI agents are so capable at continuous learning, memory management, and proactive work that the cognitive and economic cost of restarting them for each task outweighs the benefit of turning them off.

Forget abstract definitions. AGI will have arrived when an agent is so effective at continuously generating value—actively performing tasks without needing to be re-prompted—that it makes economic sense to keep it running 24/7. It's a pragmatic, economic benchmark for its arrival.

Mustafa Suleiman measures AI's human-level performance by its practical outputs. He cites an AI's ability to create a daily briefing summary that is superior to what his human chief of staff can produce as a concrete example of achieving human-level performance in a specific, valuable task.

Benchmarks like GDPVal show models like GPT-4 consistently outperform human experts on professional tasks, meeting the practical definition of AGI for knowledge work. The public discourse, however, has prematurely shifted the goalposts to sci-fi concepts of Artificial Superintelligence (ASI), obscuring the revolution already underway.

Instead of abstract technical benchmarks, the true test of an AI agent's capability is its ability to handle complex, real-world tasks. A "Mom Index" would involve challenges like logging into school software and coordinating schedules, a far better measure of practical, general intelligence.

Cutting through abstract definitions, Quora CEO Adam D'Angelo offers a practical benchmark for AGI: an AI that can perform any job a typical human can do remotely. This anchors the concept to tangible economic impact, providing a more useful milestone than philosophical debates on consciousness.

Shane Legg proposes "Minimal AGI" is achieved when an AI can perform the cognitive tasks a typical person can. It's not about matching Einstein, but about no longer failing at tasks we'd expect an average human to complete. This sets a more concrete and achievable initial benchmark for the field.

A practical definition of AGI is its capacity to function as a 'drop-in remote worker,' fully substituting for a human on long-horizon tasks. Today's AI, despite genius-level abilities in narrow domains, fails this test because it cannot reliably string together multiple tasks over extended periods, highlighting the 'jagged frontier' of its abilities.

Artificial General Intelligence Effectively Arrives When Models Match Routine B-Player Employees | RiffOn