Trending AI Posts & Hot Research Catchup - Page 10
What JWST’s Early Metal Outflows and Hubble’s Weak Stellar Winds Mean for Galaxy Evolution
16 sources0Within roughly 520–650 million years of the Big Bang, the three galaxies studied with JWST already contained enriched gas.
ChatGPT Pro Max leak: What the rumored $500 plan might offer
18 sources0Unreleased ChatGPT subscription references point to a possible $500 per month Pro Max plan offering “Fastest Work and Codex.” OpenAI has not announced it, and its speed, usage limits and launch date remain u...
Huawei’s Ascend CANN Community: Growth Claims and the Production Test
21 sources0Huawei says CANN has more than 5,200 monthly active developers, 61% of them external.
WeKnora Explained: Enterprise RAG, Agent Tools, Auto-Wiki, and the DeepSeek Harness Plugin
25 sources0WeKnora uses one enterprise knowledge layer for RAG answers, multi step agent tasks, and an interlinked Wiki.
What does the D-Robotics–GigaAI partnership aim to deliver by adapting the open-weight GigaBrain-0.7 embodied model to the 560-TOPS-INT8 Sun
17 sources0D-Robotics and GigaAI aim to turn GigaBrain-0.7 and the Sunrise S600 into a reproducible, affordable on-device robotics stack—not simply demonstrate that a model runs on a chip. The proposed design links wor...
China’s Memory Gap With South Korea: Why CXMT’s HBM Yield Matters More Than LPDDR6
21 sources0A South Korean industry association estimate puts China about three years behind in HBM, two in DRAM and one in NAND.
Huawei’s 2035 AI-Agent Forecast: 900 Billion Agents and a Ten-Part Infrastructure Plan
18 sources0Huawei envisions 900 billion AI agents by 2035, with global annual token consumption rising 100,000 fold and agents generating more than 90% of AI token traffic.
APXInf aims to make robot AI faster on the device—but its benchmark needs context
19 sources0APXInf is an open source inference engine developed by Infinigence AI with Tsinghua University and Shanghai Jiao Tong University to run embodied AI models near a robot’s sensors and actuators.
PhysBrain 1.5 Explained: One Model for Robot Understanding, Action and Prediction
17 sources0PhysBrain 1.5 uses one autoregressive model to produce spatial answers, robot motion and predicted future states.
Huawei’s Atlas 960E SuperPoD: 4,096 NPUs, Optical Links and a Faster 2027 Roadmap
19 sources0Huawei unveiled the liquid cooled Atlas 960E SuperPoD, claiming it can connect up to 4,096 NPUs for AI workloads involving 10 trillion parameter models.
HERON-CRA: How Scalabot Combines Robot Memory, Reinforcement Learning and World-Model Rollouts
18 sources0HERON CRA combines task history memory, corrective reinforcement learning and cross robot pretraining, with a proposed learning loop using imagined rollouts.
LightNav-0: How One Robot Navigation Model Aims to Reduce Teleoperation Data
16 sources0LightNav 0 is an open sourced robot navigation model built on Qwen3 VL 4B. Light Origins says training on more than 2,000 reconstructed scenes and 4,000 hours of simulated experience helps one policy transfe...