Zhemal Khamidun
Head of AI at Alpina Digital, CPO of AlpinaGPT (~30-person engineering & AI team). Founder of Vibe Academy & Hamidun Consulting, author of the AI channel «Готовим ИИшницу» (@jemal_hamidun).
PatEx-Gym Trains AI Agents to Find Prior Art: 1M Patents and a 4B Model
Researchers released PatEx-Gym, an environment for patent-examiner agents: 1,000 tasks, 1M patents, 52K papers and a compact 4B-parameter agent.
LLM agents built a Hearthstone Battlegrounds AI in 20 days that beats the top 10 human players
AutoTavern didn't play the game itself. It built a simulator and the RL agent TavernAgent. Rating rose from 6000 to 15 000, with 80% first places versus 59% for top…
CMU Researchers Teach RL Models to Stop Padding: FORK Lifts Win Rate from 22% to 74%
FORK adds a precision objective to rubric-based RL and cures bloated, inaccurate answers: on HealthBench, win rate against the base model rose from 22.1% to 74.1%.
Stanford PhD Derives the Transformer From Scratch, Starting With One Memory Problem
Zhejiang University and Peking University Propose SPADER for Multi-Answer QA
Xiaohongshu and Fudan University Build DR-Rubric for RL on Open-Ended Tasks
OpenAI exposes three Aurora models in its API: red, green and blue alpha
DynStream from Beijing builds a dynamic 3D scene from video without camera poses
Peking University, VGI Labs and the University of California present DynStream: feed-forward Gaussian reconstruction of moving scenes from streaming video, with no…
PhAI Labs: Claude Opus 5 Ranks Last in Scientific Hypothesis Test With 2.3%
PhAI Labs checked hypotheses from five model groups against later publications. Claude Opus 5 scored 2.3% and finished last; GPT got 2.5%. Here is what is behind…
Zhejiang University Publishes First Survey of Egocentric AI for Smart Glasses
Tongyan Lingzhi moves AI personalization from prompts into the model layer itself
Baidu, Tsinghua and Renmin Open-Source LoopVL, Teaching a Loop Transformer to See
GPT-6-Astra Maxes Out Robotics Benchmarks but Stumbles on Real Robots
LightSprint Raises $6.1M for a Multi-Agent Product-to-Engineering Loop
Startup LightSprint has raised $6.1M to build a system of multiple AI agents that aims to connect product teams and engineering without long approval cycles.
Cohere Proposes Making Duplicate AI Agents Compete for a Place in the Company
Cohere examines enterprise "agent sprawl": registries and reuse help, but duplicates need not be banned. They can be compared on reliability, tokens and latency.