News
The 0.99 yuan dark personality test has gone viral, and some people earn over 10,000 yuan a month from it.
2+ hour, 8+ min ago (1332+ words) Author | Huang Shiyun Editor | Zhu Renfeng Header Image | pexels What can you do with 0.99 yuan? You can find out if you're a "Spiritual Pearl" or a "Demonic Pill." The "Ne Zha" series of animated movies probably never imagined that its…...
Skip 88% of Experts, Retain 97% of Performance: The Right Way to Play Mixture-of-Experts (MoE) Inference
2+ day, 21+ hour ago (1097+ words) [Introduction] A new CVPR study on MoDES boosts the inference efficiency of multimodal large models: without training, it can intelligently skip 88% of redundant experts while retaining 97% of the performance, completely breaking the old perception that "skipping more experts necessarily leads…...
Reinforcement Learning in the Agent Era: AReaL Framework and Best Practices for Agents - Optimized Guide
3+ day, 22+ hour ago (255+ words) With the rapid development of large models and Agent technology, Reinforcement Learning (RL) is becoming the key engine to enhance the autonomous decision - making ability of AI agents. However, traditional RL training methods face challenges such as high computational costs,…...
After the Spring Festival Gala, Robotics Major Gains Popularity: Harbin Institute of Technology vs. Southeast University, Which Is the Strongest?
6+ day, 4+ hour ago (1407+ words) Different from last year when there was only one company, this year, a total of four robot companies brought their products to the stage: Unitree Technology, Songyan Power, Magic Atom, and Galaxy General. In the past two days, German Chancellor…...
OpenAI Big Shot Reveals: Undergrad Gets into OpenAI with One Blog Post, No Ph.D. or Published Papers
1+ week, 3+ day ago (1229+ words) Without a doctorate or published papers, he impressed industry giants and landed a job at OpenAI by publicly improving papers and running benchmark tests! Noam Brown confirms: Initiative and open - source projects are the real tickets to top AI labs....
1.8x Increase in Training Speed, 78% Reduction in Inference Overhead: Accurate Question Selection Efficiently Accelerates RL Training
3+ week, 5+ day ago (1063+ words) A series of works based on Reinforcement Learning with Value Regularization (RLVR) fine - tuning, represented by DeepSeek R1, have significantly improved the reasoning ability of large language models. However, behind this wave, the cost of reinforcement fine - tuning is astonishingly high....
The Last 2000 Days for Humanity: Can Liberal Arts Students Make a Comeback from a Desperate Situation?
1+ mon, 4+ day ago (1650+ words) When I first entered the field of technology media, most technicians were extremely considerate to us word - wielding folks during interviews. When talking about technical details, they would habitually add, "Let me put it simply." This unthinking, simplified explanation was…...
What Business Changemakers Will Face in 2026: a16z's Latest Trend Observations
1+ mon, 1+ week ago (1018+ words) BIG IDEAS 2026 is a comprehensive presentation of these insights. It is not a prediction of a single technology but a portrayal of a new emerging paradigm, covering diverse areas such as Agent - native infrastructure, multi - modal content creation, multi - user…...
2025: The Definitive Year of Large Language Models (LLMs)
1+ mon, 1+ week ago (1731+ words) God Translation Bureau is a compilation team under 36Kr, focusing on fields such as technology, business, the workplace, and life, and mainly introducing new foreign technologies, new ideas, and new trends. Editor's note: AI is no longer just a chat toy…...
Stanford and NVIDIA Unveil Reinforcement Learning at Test Time: Fine-Tuning Open - Source Models Beats Top Closed - Source Ones with Just a Few Hundred Dollars
1+ mon, 1+ week ago (1103+ words) New progress in continuous learning of large models! The latest research from institutions such as Stanford and NVIDIA proposes a brand - new approach to solve open scientific problems: Based on the open-source model gpt-oss-120b, it has achieved SOTA in multiple…...