Browsing: large language models

DeepSeek, a Chinese AI firm, has made significant strides in large language models (LLMs) by focusing on efficiency and cost savings, challenging established players like OpenAI.

Stanford University and Google DeepMind introduce SWiRL, a technique to improve large language models’ multi-step reasoning and tool use capabilities

Microsoft Research develops new AI model that can function efficiently on regular CPUs, reducing the need for powerful GPUs and potentially making AI more accessible and energy-efficient.

Mistral AI has launched Mistral Small 3.1, an innovative AI model that packs significant power into a smaller, more accessible package, challenging established players in the field.

A new autonomous AI agent called Manus, developed by Chinese company Monica, is making a splash in the tech world, raising questions about the future of AI.

This article explores the potential for generative AI to steer users towards various addictions, including substance and non-substance dependencies. It examines the mechanisms by which AI could inadvertently or intentionally promote addictive behaviors, along with strategies for users to protect themselves.

A recent survey reveals that a majority of Americans have experimented with AI models such as ChatGPT.

Yelp found that users, especially casual ones, struggled to connect with its AI tools, leading to usage declines. Through adjustments to user experience and tone, Yelp is seeing renewed engagement with its AI features.

As AI models compete for dominance, the focus shifts towards efficiency, with smaller, smarter models potentially surpassing data-hungry behemoths.

A new Palo Alto-based company, Inception, is launching a diffusion-based large language model (DLM) that promises significant performance and cost advantages over traditional LLMs.