A lot of what gets sent in long contexts is just noise: stale history, repeated reasoning, intermediary thinking tokens, unrelated discussions and other low-value baggage.
RESEARCH
-
Context Compaction Superiority Over Larger Context Windows
By
–
Better context compaction > bigger context windows Change my mind
-
Creating Humble AI Systems According to MIT Research
By
–
How to create “humble” #AI
by Anne Trafton @MIT Learn more: https://
bit.ly/47q4fh0 #ArtificialIntelligence #MachineLearning #ML -
New Method Boosts LLM Training Efficiency at MIT
By
–
New method could increase #LLM training efficiency
by @aczewe @MIT Learn more: https://
bit.ly/3OObtES #GenerativeAI #ArtificialIntelligence #MachineLearning #ML -
Alignment: The Key to Effective Human-AI Collaboration
By
–
Alignment is the Secret to Human-#AI Teamwork
by Lina Zeldovich @NeuroscienceNew Learn more: https://
bit.ly/3PBLEs4 #MachineLearning #ArtificialIntelligence #ML -
Unitree Open-Sources Dataset for Humanoid Robot Teleoperation
By
–
🔥 Dancing and Kung Fu are just the warm up, #Unitree is building real world AI.
— Amitav Bhattacharjee (@bamitav) 28 mars 2026
Unitree Robotics just released open-sources the UnifoLM-WBT-Dataset: a high-quality, real-world dataset for whole-body teleoperation (#WBT) of humanoid robots in open environments.#OpenSource… pic.twitter.com/wk2mINZk1HDancing and Kung Fu are just the warm up, #Unitree is building real world AI. Unitree Robotics just released open-sources the UnifoLM-WBT-Dataset: a high-quality, real-world dataset for whole-body teleoperation (#WBT) of humanoid robots in open environments. #OpenSource
-
RL Training for Distributional Reasoning in Language Models
By
–
"Reaching Beyond the Mode: RL for Distributional Reasoning in Language Models" Instead of standard RL post-training collapsing an LLM toward one dominant answer, this paper shows you can train it to produce a set of plausible answers in a single pass. This is important because
-
Language Models Drive Novel Scientific Discovery Beyond Benchmarks
By
–
LLMs aren't just chatbots, they can also search for novel discoveries! In this AI4Science talk, Yuanqi Du (
@YuanqiD
) walks through a shift in how to think about language models in science. Instead of asking whether they “understand” science through benchmarks or exams, the work -
DeepMind’s Self-Improving AI Conquers Table Tennis
By
–
DeepMind’s Self-Improving Table Tennis #AI Takes on the Game
— Ronald van Loon (@Ronald_vanLoon) 28 mars 2026
via @ZappyZappy7#Robotics #MachineLearning #ArtificialIntelligence #ML #Innovation pic.twitter.com/h53hmLaQOGDeepMind’s Self-Improving Table Tennis #AI Takes on the Game
via @ZappyZappy7 #Robotics #MachineLearning #ArtificialIntelligence #ML #Innovation -
AI’s Potential to Enhance Public Health and Global Well-being
By
–
Can #AI improve public health and support global well-being?
by @antgrasso #Healthcare #HealthTech #Tech #Technology