Newsletter Archives

• LLMs can’t reason

  Posted on December 30, 2024 at 02:44 CST by Michael Covington • Comment in the Forums

  AI

  

  By Michael A. Covington

  The word is out — large language models, systems like ChatGPT, can’t reason.

  That’s a problem, because reasoning is what we normally expect computers to do. They’re not just copying machines. They’re supposed to compute things. We knew already that chatbots were prone to “hallucinations” and, more insidiously, to presenting wrong answers confidently as facts.

  But now, researchers at Apple have shown that large language models (LLMs) often fail on mathematical word problems.

  Read the full story in our Plus Newsletter (21.53.0, 2024-12-30).

  AI AI, Artificial Intelligence, Large language model, Machine Learning, Math Word Problems, Neural Networks, Neuroscience, Newsletters, Reasoning

• Microsoft 365 Copilot announced

  Posted on March 27, 2023 at 02:44 CDT by Will Fastie • Comment in the Forums

  MICROSOFT NEWS

  

  By Will Fastie

  Another shoe has dropped in Microsoft’s quest to build generative AI into just about everything.

  On March 16, Microsoft announced a new feature named Copilot, which will be provided with Microsoft 365 accounts. Copilot uses generative artificial intelligence, a large language model (LLM), and the Microsoft Graph to provide services based on natural language prompts. Examples of such prompting are evident in the controversial BingAI, but Microsoft emphasized its commitment to “responsible AI” for Copilot.

  Read the full story in our Plus Newsletter (20.13.0, 2023-03-27).

  Microsoft News Artificial Intelligence, Bing AI, Copilot, Generative AI, Large language model, Microsoft 365, Microsoft Graph, Newsletters