Tencent’s new AI technique teaches language models ‘parallel thinking’

The Parallel-R1 framework uses reinforcement learning to teach models how to explore multiple reasoning paths at once, ...

DeepSeek claims training R1 model cost less than $300,000; acknowledges A100 use

DeepSeek (DEEPSEEK) revealed that its R1 model was trained at a much lower cost than what U.S. competitors have seen, ...
Opinion
Harvard Graduate School of EducationOpinion

Askwith Education Forum Dives Deep into AI In Education

“Thinking in an AI-Augmented World” featured Professor Howard Gardner in conversation with award-winning international law scholar and founder of Dragonfly Thinking Anthea Roberts, who shared how she ...

We Finally Know How Much It Cost to Train China’s Astonishing DeepSeek Model

DeepSeek found that it could improve the reasoning and outputs of its model simply by incentivizing it to perform a trial-and ...

DeepMind AI safety report explores the perils of “misaligned” AI

DeepMind's safety framework is based on so-called "critical capability levels" (CCLs). These are essentially risk assessment rubrics that aim to measure an AI model's capabilities and define the point ...