The original version of this story appeared in Quanta Magazine. Large language models work well because they’re so large. The latest models from OpenAI, Meta, and DeepSeek use hundreds of billions of ...
This article predates the current owners of Science Blog. It has not been reviewed under the current editorial process and is kept as part of the archive. A new approach developed by MIT CSAIL ...
This article is long. However, to make it worth revisiting, I have provided three guideposts. Guidepost 1: Confidence Labels ...
When Liquid AI, a startup founded by MIT computer scientists in 2023, introduced its Liquid Foundation Models series 2 (LFM2) in July 2025, the pitch was straightforward: Deliver the fastest on-device ...
Microsoft just released its latest small language model that can operate directly on the user’s computer. If you haven’t followed the AI industry closely, you might be asking: what exactly is a small ...