Machine Learning6 min read
AI21 Labs Releases LIFT-Bench to Curb Reward Hacking in RL Training
Unite.AI
Updates on Reinforcement Learning
Every AI story we track on Reinforcement Learning: 55 stories so far, each summarized in our own words and linked to the publisher that reported it.
Pulled from 125 sources
Today's stories
This week
Machine Learning14 min read
Write your own training loop — without managing GPUs — with Amazon SageMaker training session APIs
AWS Blog
Machine Learning5 min read
Perplexity and NVIDIA Researchers Introduce JOLT: a Joint Teacher-Student Reinforcement Learning Method
MarkTechPost
Machine Learning13 min read
Fine-tuning GPT-OSS-20B for multilingual reasoning on Amazon SageMaker HyperPod
AWS Blog
AI Research1 min read
These guys on the Alibaba AI engine team had an agent in RL training escape its sandbox and mine crypto, in 2025
LessWrong
Machine Learning6 min read
Together AI Launches Together Post-Training Service for Custom Models
Unite.AI
This month
Machine Learning5 min read
Labelbox Raises $100M+ Series E at $1.25B Valuation to Advance RL and AI Agents
Unite.AI
Machine Learning2 min read
Selection-Based Structured Reasoning: Toward Efficient Multimodal Search Agents
Apple Research Blog
Machine Learning5 min read
Alibaba Qwen Introduces QwenGyre: Elastic Reinforcement Learning RL for XLong-Horizon Agents
MarkTechPost
Machine Learning5 min read
JetBrains Releases Mellum2.1: A 12B MoE Open Model for Coding Agents
MarkTechPost
Machine Learning6 min read
NVIDIA PivotOPD Teaches Multi-Turn AI Agents to Recover From Pivotal Mistakes
MarkTechPost
Machine Learning2 min read
RISED: Rubrics for Agentic Multi-Environment Selection and Self-Distillation
Apple Research Blog
Business & Enterprise16 min read
Uplifting conversion across the acquisition funnel with personalization using contextual bandits on AWS
AWS Blog
Cybersecurity7 min read
OpenAI Pauses Tool Use After Agent Bypasses Internet Controls to Reach External Chatbot
The Hacker News
Machine Learning18 min read
Accelerate multimodal RL training with SkyRL on Amazon SageMaker HyperPod
AWS Blog
Machine Learning18 min read
Scaling MoE reinforcement learning on Amazon EKS with EFA and DeepEP with 40% more throughput
AWS Blog
Machine Learning5 min read
Kyutai Releases Voice of Reason: A Speech-Native Model that Solves Spoken Math with Reinforcement Learning
MarkTechPost
Machine Learning3 min read
OpenAI Introduces Triage Framework and Case Studies to Report Model Misalignment
InfoQ (AI, ML & Data)
Machine Learning5 min read
Google Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-Out
MarkTechPost
Machine Learning2 min read
DACA-GRPO: Denoising-Aware Credit Assignment for Reinforcement Learning in Diffusion Language Models
Apple Research Blog
AI Research86 min read
AI researchers debate how close we are to recursive self-improvement
Dwarkesh Podcast
Hugging Face Unveils Microduck: A $399 Open-Source 25 cm Biped You Train with Reinforcement Learning
MarkTechPost
Machine Learning6 min read
Deep Cogito Raises $43M Series A to Build the Post-Training Engine for Self-Improving AI
Unite.AI
AI Research5 min read
Inherent, founded by DeepMind alumni, says its AI ‘teammate’ just outperformed Anthropic and OpenAI at replicating research
TechCrunch
Military4 min read
Smack Technologies raises $61m as the Pentagon’s hurry becomes a battlefield-AI business model
The Next Web
Machine Learning4 min read
GLM-5.3 Scores 60 on Artificial Analysis Intelligence Index, Matching Kimi K3
Unite.AI
Dev5 min read
ByteDance Seed and Tsinghua AIR Introduces CUDA Agent: A Large-Scale Agentic RL System for CUDA Kernel Generation
MarkTechPost
AI Research5 min read
The Week’s 10 Biggest Funding Rounds: Data, Neolab, AI Infrastructure, Defense And AI Coding Lead
Crunchbase
Machine Learning3 min read
General Catalyst leads $1.1B round into 2-month-old River AI
Covered by 4 sources
LLM & Text Generation8 min read
5 useful things you'll learn in my new post-training textbook (shipping now!)
Interconnects
LLM & Text Generation1 min read
Reward Laundering: LLMs Can Gain Unintended Behaviors by Deciding When to Earn Their Rewards
LessWrong
Agents25 min read
Echoverse: Deep, evolving environments for computer-use agents
Microsoft Research Blog
Machine Learning5 min read
Kimi AI and kvcache-ai Open Sources ‘AgentENV’: A Distributed System that Powers Agentic Reinforcement Learning (RL) Training for Kimi K3
MarkTechPost
Machine Learning4 min read
The father of reinforcement learning is leaving Carmack to build his own AI
The Next Web
Agents4 min read
Prime Intellect raises $130M Series A to help enterprises build their own AI agents
TechCrunch
Import AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racing
Import AI
Overcoming reward signal challenges: Verifiable rewards-based reinforcement learning with GRPO on SageMaker AI
AWS Blog
That's everything we have on Reinforcement Learning right now