ResearchScore smoothing during training explains diffusion models' tendency to interpolateA paper presented at ICLR 2026 argues that the apparent "creativity" of diffusion models—their ability to produce novel images or molecular structures—follows from how neural-network training smooths the learned score function.4 min read
ResearchPublic AI Models Rapidly Solved Long‑standing Theoretical Problems and Self‑Verified Their WorkIn July 2026, OpenAI reported that GPT-5.6 Sol Ultra solved the Cycle Double Cover conjecture in under an hour by splitting into 64 agents that critiqued each other's proofs.2 min read
ResearchApple in Early Talks with PrismML on Running Compressed LLMs Locally on iPhoneApple is in preliminary discussions with PrismML, a Caltech spin‑off backed by Khosla Ventures, which says it has compressed large language models so they can run on iPhone 15 and newer devices.3 min read
ResearchNew benchmark Real World VoiceEQ measures human-quality aspects of voice AIReal World VoiceEQ is a new benchmark designed to assess how well voice AI systems capture and act on acoustic cues that transcripts omit—tone, emotion, speaker identity and background context.5 min read
ResearchPractical lessons from the Nemotron Model Reasoning Challenge: five ways teams improved reasoning with open modelsThe NVIDIA Nemotron Model Reasoning Challenge on Kaggle drew over 5,000 participants in roughly 4,000 teams, producing thousands of submissions and more than 1,000 discussion posts.6 min read
ResearchNVIDIA’s Ising Decoder Speeds and Improves Color Code Decoding for Quantum Error CorrectionNVIDIA released Ising Decoder ColorCode 1 Fast, a 3D CNN pre-decoder for triangular color codes that reportedly reduces logical error rates and shortens decoding time compared with the open-source Chromobius decoder.3 min read
ResearchAnthropic identifies a hidden 'J-space' of internal tokens inside large language modelsAnthropic reports finding a previously unseen internal representation space—called the J-space—in its Claude models, populated by tokens that do not appear in outputs but seem to track reasoning steps and states.3 min read
ResearchClaude's expressions of value vary by language: warmth vs. strictnessClaude's linguistic behavior changes depending on the language of the conversation; the difference is most noticeable on the warmth versus strictness axis.1 min read
ResearchResearch into the causes and controllability of the language model Claude's value expressionsThe value manifestations of the language model Claude affect everyday conversations, but the developers do not understand why these values change.1 min read
ResearchAgent-as-a-Router and ACRouter: a dynamic, feedback-driven approach to model routingResearchers propose Agent-as-a-Router, an open-source paradigm that treats the model router as a learning agent using a Context-Action-Feedback loop to accumulate execution-grounded memory.6 min read
ResearchHungarian AI camera system monitors preterm infants' movements and positionsResearchers in Hungary developed a camera-based monitoring system that uses deep learning to track the spontaneous movements and body positions of preterm infants in incubators.2 min read
ResearchPositive assessment of the GPT-5.6 family of modelsThe article's author evaluated the GPT-5.6 family of models, finding that they deliver excellent performance.1 min read
