A large team of tech nostalgia enthusiasts have made a PiDP-10, a replica of the PDP-10 mainframe computer first launched by the Digital Equipment Corporation in 1966.
An article by Phillip Morrison exploring the concept of machine thinking in the age of computers. The author discusses the recurring question of why life is different from non-life, with each new generation of machines that simulate aspects of life.
This paper introduces Cross-Layer Attention (CLA), an extension of Multi-Query Attention (MQA) and Grouped-Query Attention (GQA) for reducing the size of the key-value cache in transformer-based autoregressive large language models (LLMs). The authors demonstrate that CLA can reduce the cache size by another 2x while maintaining nearly the same accuracy as unmodified MQA, enabling inference with longer sequence lengths and larger batch sizes.
MIT CSAIL researchers have developed three neurosymbolic frameworks - LILO, Ada, and LGA - that use natural language to help large language models (LLMs) build better abstractions for coding, AI planning, and robotics tasks.
This article discusses the MIT Artificial Intelligence (AI) Lab's 'Tourist Policy' and how it impacted students' access to its resources. As a high school student in Maryland, the author shares their experience of using the lab's PDP-10s over the ARPANET and how it inspired them to learn and contribute to the MIT community.
Bookmarks on this page are managed by an admin user. 