
A framework organizes AI beliefs into three stages: recognition of current AI capabilities, belief that AI will become much more powerful, and belief that AI will eventually exceed human performance at nearly all tasks within our lifetimes.
The author and many frontier AI lab employees hold the strongest belief, while the majority of people remain unaware of what current AI systems can already do.
What happened
A framework categorizes AI beliefs into three "pills": AI pilled (current capabilities exist), AGI pilled (AI will do much more), and ASI pilled (AI will eventually outperform humans at nearly everything within our lifetimes). The author states they and many frontier lab employees and labs themselves are ASI pilled.
Why it matters
Most people have not adopted even the first pill and lack awareness of what current frontier AI can do, such as coding agents. The framework suggests sincere disagreements about AI typically stem from disagreements about future capabilities rather than disagreement about present facts.
What to watch
The author notes the majority of people remain unaware of frontier AI's current capabilities and have only used ChatGPT for basic tasks, indicating a significant gap between what the technology can do and public understanding of it.
The author presents a framework for understanding disagreements about artificial intelligence by organizing beliefs into three distinct positions, called "pills." The first, AI pilled, represents the most basic acknowledgment that artificial intelligence exists and can perform the tasks it currently performs. The second, AGI pilled, represents the belief that AI will become capable of doing much more than it presently can. The third, ASI pilled, represents the strongest position: that AI will eventually be able to perform approximately all tasks better than humans within our natural lifetimes. The author explicitly states they adopt the ASI position. They further note that a large percentage of employees at frontier AI labs—organizations developing cutting-edge AI systems—are also ASI pilled, and that the labs themselves operate from this perspective. The article then shifts to describing those who have not adopted any of these frameworks. The author observes unpilled people (those who have taken zero pills) are widespread, making up the majority of the population. Most people remain unaware of what frontier AI systems are capable of doing and lack knowledge of specific applications such as coding agents. Their exposure to AI has typically been limited to ChatGPT used for minor, nontechnical tasks, and these interactions have left them with outdated impressions of AI capabilities.
The article presents a three-level framework for categorizing positions on artificial intelligence development, arguing that sincere disagreements about AI typically reflect disagreement about future capabilities rather than present reality. The framework distinguishes between acknowledging current AI abilities (AI pilled), accepting that AI will become significantly more capable (AGI pilled), and believing that AI will eventually match or exceed human performance across nearly all domains within the current generation (ASI pilled). The author identifies themselves and a large portion of frontier AI lab personnel as holding the strongest belief (ASI pilled), suggesting this represents the consensus view among those most directly involved in developing advanced AI systems. The article contends that there exists a substantial gap between technical capability and public awareness: most people lack familiarity with what frontier AI systems can currently accomplish and have interacted with AI only through basic use cases.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
Running a 122-billion-parameter model on three RTX 3090 GPUs with a 256K-token context, the author's AI agent…

NVIDIA and partners released multiple open-source AI models optimized for local execution throughout August, i…

Major technology companies are advocating for a new standardized framework to report incidents involving AI ag…

Anthropic and Cisco integrated Cisco AI Defense with Claude Enterprise's inference hooks, a security layer tha…

Honeywell Technologies is hosting a webinar to explore agentic AI (autonomous AI systems that can act independ…

Target appointed Chandhu Nair as its first chief artificial intelligence officer and senior vice president on…

The AI news that matters, in one minute each morning.
Sign up free