S

Scaled Cognition

AI · Berkeley, United States · Founded 2023

Scaled Cognition, founded in 2023 by human-computer interaction researcher Alan Kay, is developing artificial intelligence systems with human-like common sense to enable more effective collaboration between humans and AI. Kay, a luminary in the field of personal computing and object-oriented programming, brings decades of experience from pioneering work at Xerox PARC, Apple, and as a Turing Award laureate, shaping the company's ambitious vision. The company is focused on building what it terms 'Model-Based Agents,' a departure from the current large language models (LLMs). Instead of simply predicting the next word, Scaled Cognition's AI aims to construct an internal model of the world to understand context, reason about problems, and learn from interactions in a manner more analogous to human cognition. This approach is designed to overcome the limitations of existing AI, such as hallucinations and a lack of genuine understanding, by imbuing the system with a framework for common sense. The primary offering is an AI system architected for 'agentic applications,' where the AI can act as a proactive partner to human users. This system is intended to serve developers and organizations looking to build applications that require a deeper level of reasoning and reliable performance. The business model appears to be centered on providing access to this specialized AI, with a waitlist available for early access, suggesting a future B2B or API-based revenue stream. Keywords: agentic AI, model-based agents, common sense reasoning, human-computer interaction, artificial general intelligence, Alan Kay, cognitive architecture, AI reasoning, contextual understanding, AI safety