10 primary articles · 3 secondary mentions
Problem — This work addresses the gap in understanding the cognitive and epistemic implications of AI systems, particularly through the lens of cognitive colonization. While existing frameworks like Tri-System Theory...
Problem The paper addresses the insufficient characterization of the alignment between large language models (LLMs) and the neural mechanisms of human higher-order cognition, particularly in the context of deductive reasoning....
Cognition’s Scott Wu, the mind behind Devin—one of the leading AI coding agents—asserts that these technologies are not intended to replace human programmers. As AI continues to evolve and integrate...
Problem This paper addresses the limitations of current large language models (LLMs) in reasoning tasks, particularly the inefficiencies associated with autoregressive generation during test-time computation. The authors argue that this...
Cognition, the company behind the AI coding agent Devin, has successfully raised over $1 billion, propelling its valuation to over $26 billion in less than nine months. This significant funding...
Cognition, an AI coding startup, has successfully raised $1 billion in funding, achieving a staggering pre-money valuation of $25 billion. This significant financial milestone comes as the company reports an...
Problem This paper addresses the gap in the governance and lifecycle management of agent-generated artifacts within multi-agent systems, particularly focusing on their evolution as persistent runtime capabilities. Prior works, such...
Problem This paper addresses a significant gap in the evaluation of Vision-Language Models (VLMs) concerning their understanding of camera motion, which is crucial for spatial cognition. Existing spatial VLMs demonstrate...
Problem This paper addresses a gap in the understanding of how large language models (LLMs) process contextual cues, particularly in high-stakes decision-making scenarios. The authors explore the phenomenon of inattentional...
Problem This preprint addresses the gap in understanding how large language models (LLMs) can efficiently approximate global activation patterns for long-context understanding. Existing models struggle with the computational burden of...
