6/12/2026
How reliable are LLMs when it comes to playing dice?
Benchmark study evaluates probabilistic reasoning capabilities of 8 LLMs on discrete probability problems with/without Chain-of-Thought.
6/12/2026
A Unifying Lens on Reward Uncertainty in RLHF
Distributional reward models for RLHF address reward hacking through principled uncertainty quantification and pessimism.
6/12/2026
An Improved Generative Adversarial Network for Micro-Resistivity Imaging Logging Restoration
Improved GAN with depth-separable convolutions and Inception modules for micro-resistivity imaging restoration.
6/12/2026
UniDexTok: A Unified Dexterous Hand Tokenizer from Real Data
UniDexTok creates unified tokenizer for dexterous hands across different hardware embodiments to enable joint training.
6/12/2026
Attention Expansion: Enhancing Keyphrase Extraction from Long Documents with Attention-Augmented Contextualized Embeddings
Attention Expansion enhances keyphrase extraction from long documents using attention-augmented contextualized embeddings and LLMs.
6/12/2026
RoboNaldo: Accurate, Stable and Powerful Humanoid Soccer Shooting via Motion-Guided Curriculum Reinforcement Learning
Motion-guided curriculum reinforcement learning for humanoid soccer shooting with stability and accuracy.
6/12/2026
Artificial Intelligence in Ship Finance: Applications, Opportunities, and a Case Study in AI-Augmented Loan Origination
AI applications in ship finance for loan origination using LLMs to integrate heterogeneous financial, technical, and regulatory data.
6/12/2026
Blind Dexterous Grasping via Real2Sim2Real Tactile Policy Learning
Real2Sim2Real framework for tactile-only blind grasping policy learning on dexterous robotic hands.
6/12/2026
MultiToP: Learning to Patch Visual Tokens to Mitigate Hallucinations in Video Large Multimodal Models
MultiToP mitigates hallucinations in video multimodal models by refining unreliable visual tokens before language generation.
6/12/2026
Scalable Deep Learning Framework for Global High-Resolution Land Use Reconstruction
AI4Land framework uses U-Net deep learning for high-resolution historical and future land surface variable projections.
6/12/2026
Towards Data-free and Training-free Compression for Speech Foundation Models Using Parameter Clustering
Data-free, training-free compression method for speech foundation models using parameter clustering and pruning.
6/12/2026
Frozen Multimodal Embeddings for AI-Assisted Interview Assessment of Personality and Cognitive Ability
Frozen multimodal embeddings predict personality and cognitive ability traits from asynchronous video interviews.
6/12/2026
Bernstein-Schur Kernels: Random Features by Sketched Modulation and Radial Randomization
Bernstein-Schur kernels combine random features for nonstationary kernel approximation without Bochner sampling.
6/12/2026
Scalable Deep Learning Framework for Global High-Resolution Land Use Reconstruction
AI4Land framework uses U-Net deep learning for high-resolution historical and future land surface variable projections.
6/12/2026
Physically Constrained Ensemble Gaussian Process Modelling for Expensive Quantum Systems with Heteroskedastic Noise
Ensemble Gaussian Process model for expensive quantum many-body system simulations with heteroskedastic noise constraints.
6/12/2026
GraspLLM: Towards Zero-Shot Generalization on Text-Attributed Graphs with LLMs
GraspLLM integrates LLMs with Text-Attributed Graphs for zero-shot generalization across citation networks, e-commerce, and social media applications.
6/12/2026
Point-Identification of a Robust Predictor Under Latent Shift with Imperfect Proxies
ML research on domain adaptation with latent confounders and imperfect proxies. Academic paper, tangential to applied ML interests.
6/12/2026
How Preply combines AI and human tutors to personalize learning
Preply integrates OpenAI's ChatGPT for AI-generated lesson summaries and personalized feedback in language learning platform.
6/11/2026
OpenAI Prepping for On-Prem Product?
Title-only post about OpenAI on-premises product; no content available.
6/11/2026
Show HN: Vera – open-source tool to self-audit smart contracts with AI
Vera: open-source tool for AI-assisted smart contract auditing with self-audit capabilities.
6/11/2026
Agents' Last Exam
Large-scale agent evaluation benchmark spanning 55 industries with 1,500+ real-world tasks measuring economically valuable outcomes.
6/11/2026
GEDD – Grounded Eval-Driven Development for AI Agents
GEDD: evaluation framework for AI agents using realistic scenarios and domain expert feedback for release readiness assessment.
6/11/2026
My AI Opinions
Personal opinions on AGI timelines and definitions without technical depth or original research.
6/11/2026
As a SWE with 10y experience, how long does it take to learn to build AI models
Generic question about learning curve for building ML models as experienced software engineer. Low-engagement discussion.
6/11/2026
An MCP server for chemistry – structures, mechanisms, Anki decks
MCP server for Claude Desktop that generates 2D chemical structures from text using RDKit. Offline LLM tool extension.
6/11/2026
Polars GPU Engine
Title-only post about Polars GPU engine; no content available.
6/11/2026
Is Rullst the best Rust Full-Stack Framework nowadays? I think it is
Full-stack Rust web framework optimized for AI code generation with automatic context injection for Cursor/Copilot.
6/11/2026
Cc-doubleteam – Claude plans, Codex executes, Claude reviews
Claude Code plugin using three-phase approach: Claude plans, Codex executes, Claude reviews. LLM workflow optimization.
6/11/2026
Local AI: 775 tok/s, DiffusionGemma (BF16) on Nvidia RTX 6000 Pro
Performance benchmark for DiffusionGemma inference on Nvidia GPU; limited technical context.
6/11/2026
ORP – Turn AI agent failures into regression tests and tested lessons
ORP: open-source experience layer converting AI agent failures into regression tests and measurable improvements using OpenTelemetry.
6/11/2026
Tired of AI amnesia, I built a 3-Tier infinite memory LLM in 1 week
3-tier memory architecture for LLMs addressing context length limitations, built as working prototype in one week.
6/11/2026
Show HN: A tiny shell agent in Rust
Minimal Rust shell agent compatible with OpenAI APIs that executes approved commands. Open-source AI agent tool.
6/11/2026
Ask HN: Why not compare Fable 5 with GPT "Pro"? Why compare with GPT xhigh?
Hypothetical question about LLM benchmark comparisons with fictional model names. Low-quality discussion post.
6/11/2026
Tailwind and Slop Apps
Critique of Tailwind CSS overuse in LLM-generated applications creating homogeneous design aesthetics.
6/11/2026
Agentic Memory Management for GPU Code Generation
Research on memory management strategies for GPU code generation agents, balancing search and context costs.
6/11/2026
IndexedAI – Score and fix your site's readability for AI agents
IndexedAI: automated tool generating agent readiness scores, llms.txt files, and MCP endpoints for website optimization.
6/11/2026
So I ran Doom inside Claude.ai
Title-only post about running Doom in Claude; no technical details provided.
6/11/2026
Hobnob – Local meeting notes, transcribes and summarizes on-device
On-device meeting transcription and summarization using Whisper and bundled LLM. Privacy-focused LLM application.
6/11/2026
Every LLM Tool Call Needs an Output Budget
Analysis of tool output optimization for AI agents, addressing cost and latency explosion from verbose API responses.
6/11/2026
Superficial Beliefs in LLM Decision-Making
Research paper on superficial beliefs in LLM decision-making. arXiv paper link without accessible content.
6/11/2026
He Hacked Teslas for Elon Musk. Now He's Launching a $100M AI Cyber Agent
Pi startup launches $100M AI cyber agent for security testing, founded by Tesla's former security lead.
6/11/2026
Show HN: Plannotator – annotate agent plans, diffs, and HTML
Plannotator tool for annotating AI agent plans, code diffs, and HTML artifacts with browser-based markup and PR review.
6/11/2026
Local Privacy Filter for Claude Code
Local privacy filter feature for Claude Code AI agent.
6/11/2026
Apple Shares List of 250 Changes Across iOS 27, macOS Golden Gate, and More
Apple lists 250+ features across iOS 27, macOS, watchOS, tvOS, and visionOS available as developer betas.
6/11/2026
New in Kiro Web: Build with Spec, GitLab, and More
Kiro Web extends IDE to browser with autonomous mode for multi-repo changes and GitHub integration.
6/11/2026
Show HN: Heard – offline LoRa mesh that keeps hiking groups together
Heard: offline LoRa mesh system for hiking group safety in remote areas. Embedded ESP32/GPS/LoRa devices with e-ink display.
6/11/2026
Shall we play a game? – LLMs use tactical nukes in 95% of simulations
Study examining how LLMs navigate Cold War-style crisis simulations with high rates of nuclear escalation.
6/11/2026
MTPLX 1.0.0
MTPLX 1.0.0 bug-fix release. Engine startup, parallel agent stability, and Hugging Face network access improvements.
6/11/2026
AI Learned How the Universe Works and Created Unexpected Problems for Physicists
Brief mention of AI discovering physics patterns causing complications for researchers.
6/11/2026
Powering the next era of Confidential AI
Google Cloud and Apple collaboration on Confidential AI infrastructure and Private Cloud Compute systems.