HN AkshatVirmani 3/30/2026

APIEval-20: A Benchmark for Black-Box API Test Suite Generation

APIEval-20 benchmark dataset for evaluating black-box API test suite generation using LLMs and schemas.

HN u1hcw9nx 3/30/2026

Subnautica 2 publisher's CEO used ChatGPT to fire studio head and failed

CEO used ChatGPT to terminate studio head; decision was reversed and criticized.

HN btursunbayev 3/30/2026

Show HN: NVSonar – tells why your GPU is slow, not just utilization percentage

GPU profiling tool that diagnoses performance bottlenecks beyond utilization metrics. Minimal details provided but relevant tool.

HN thiagolalvarez 3/30/2026

I built an MCP server so your agent stops picking the wrong cloud services

MCP server for AI agents to select appropriate cloud services with current pricing and compatibility data. 74 services, no API key required.

HN LionTurtle13 3/30/2026

Stanford study reveals AI vision models invent images they never see

Stanford research showing AI vision models generate images not in training data through hallucination mechanisms.

HN KnuthIsGod 3/30/2026

President Trump Gaggles with Press on Air Force One En Route

News about President Trump press interaction on Air Force One.

HN walterbell 3/30/2026

Tribe v2: Predictive Foundation Model on Human Brain Processing Complex Stimuli

TRIBE v2: Predictive AI model of human brain responses to visual, auditory, and language stimuli from neuroscience research.

HN userbinator 3/30/2026

HD Audio Driver for Windows 98SE / Me

HD Audio driver for Windows 98SE/ME systems on Intel chipsets with WDM support.

HN bthallplz 3/30/2026

Excel2r – R package that migrates Excel workbooks to standalone R scripts

R package that converts Excel workbooks to standalone R scripts with formula recreation and verification against cached values.

HN keiranflynn 3/30/2026

LLMnesia – Local-first search across your AI conversations

LLMnesia: Local-first search tool for AI conversation history across ChatGPT, Claude, Gemini, and other platforms.

HN 1vuio0pswjnm7 3/30/2026

Meta's court losses spell potential trouble for AI research, consumer safety

Analysis of Meta's legal losses and liability implications from internal social science research on platform effects.

HN newtechwiz 3/30/2026

Show HN: Free, Open-Source WhisperFlow That Just Works

WhisperFlow: Free, open-source speech-to-text tool for macOS. On-device processing, no cloud upload, no account required.

Ax Hao Li, Qiao Sun 3/30/2026

Any4D: Open-Prompt 4D Generation from Natural Language and Images

arXiv research on 4D generation from natural language and images using embodied world models. Addresses data scarcity and long-horizon video generation challenges.

Ax Zhenchao Tang, Fang Wang, Haohuai He, Jiale Zhou, Tianxu Lv, Jun Zhu, Shouzhi Chen, Minghao Yang, Yu Wang, Jiayang Wu, Yidong Song, Yaokun Li, Jiehui Huang, Dawei Huang, Zhi Song, Jianhua Yao 3/30/2026

Aligning LLMs with Biomedical Knowledge using Balanced Fine-Tuning

arXiv research proposing Balanced Fine-Tuning method for aligning LLMs with biomedical knowledge. Combines SFT and RL using confidence-weighted token optimization for scientific understanding.

Ax Daeun Lee, Subhojyoti Mukherjee, Branislav Kveton, Ryan A. Rossi, Viet Dac Lai, Seunghyun Yoon, Trung Bui, Franck Dernoncourt, Mohit Bansal 3/30/2026

StreamGaze: Gaze-Guided Temporal Reasoning and Proactive Understanding in Streaming Videos

arXiv research on streaming video understanding with gaze signal interpretation for AR applications. Evaluates multimodal LLMs on temporal reasoning with human attention signals.

Ax Woongyeong Yeo, Kangsan Kim, Jaehong Yoon, Sung Ju Hwang 3/30/2026

WorldMM: Dynamic Multimodal Memory Agent for Long Video Reasoning

arXiv research on multimodal memory architecture for long-form video understanding. Addresses context capacity and visual detail retention in hours-long videos using dynamic memory mechanisms.

Ax David Samuel, Lilja {\O}vrelid, Erik Velldal, Andrey Kutuzov 3/30/2026

Fluent Alignment with Disfluent Judges: Post-training for Lower-resource Languages

Post-training method for lower-resource languages preserving fluency when aligned by disfluent reward models, addressing preference optimization data scarcity.

Ax Ruining Li, Yuxin Yao, Chuanxia Zheng, Christian Rupprecht, Joan Lasenby, Shangzhe Wu, Andrea Vedaldi 3/30/2026

Particulate: Feed-Forward 3D Object Articulation

Feed-forward transformer model predicting 3D object articulations including parts, kinematic structure, and motion constraints for articulated object understanding.

Ax Boxin Wang, Chankyu Lee, Nayeon Lee, Sheng-Chieh Lin, Wenliang Dai, Yang Chen, Yangyi Chen, Zhuolin Yang, Zihan Liu, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping 3/30/2026

Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models

Cascaded reinforcement learning infrastructure for scaling general-purpose reasoning models, addressing heterogeneity in response lengths and verification latency.

Ax Wentao Guo, Mayank Mishra, Xinle Cheng, Ion Stoica, Tri Dao 3/30/2026

SonicMoE: Accelerating MoE with IO and Tile-aware Optimizations

SonicMoE optimizes Mixture of Experts model inference through IO and tile-aware techniques, accelerating high-sparsity MoE architectures for language models.

Ax Zhijie Zhong, Zhiwen Yu, Pengyu Li, Jianming Lv, C. L. Philip Chen, Min Chen 3/30/2026

PathFinder: Advancing Path Loss Prediction for Single-to-Multi-Transmitter Scenario

Deep learning method for radio path loss prediction in multi-transmitter 5G scenarios, addressing distribution shifts and environmental generalization.

Ax David Samuel, Lucas Georges Gabriel Charpentier 3/30/2026

Dual-objective Language Models: Training Efficiency Without Overfitting

Dual-objective language model combining autoregressive and masked-diffusion training without architectural changes, improving efficiency and reducing overfitting.

Ax Pengyu Wang, Shuchang Ye, Usman Naseem, Jinman Kim 3/30/2026

MRG-R1: Reinforcement Learning for Clinically Aligned Medical Report Generation

Medical report generation using reinforcement learning with clinical alignment objectives, improving correctness over token-level likelihood training approaches.

Ax Sara Papi, Javier Garcia Gilabert, Zachary Hopton, Vil\'em Zouhar, Carlos Escolano, Gerard I. G\'allego, Jorge Iranzo-S\'anchez, Ahrii Kim, Dominik Mach\'a\v{c}ek, Patricia Schmidtova, Maike Z\"ufle 3/30/2026

Hearing to Translate: The Effectiveness of Speech Modality Integration into LLMs

Study comparing SpeechLLMs that directly process speech for translation against cascaded transcription pipelines, evaluating speech modality integration effectiveness.

Ax Matthew Thompson 3/30/2026

The Dual-State Architecture for Reliable LLM Agents

Dual-State Architecture formalizes execution primitives coupling stochastic LLM generation with deterministic verification guards for reliable code generation agents.

Ax Subeen Lee, Siyeong Lee, Namil Kim, Jaesik Choi 3/30/2026

RoAD Benchmark: How LiDAR Models Fail under Coupled Domain Shifts and Label Evolution

Benchmark evaluating LiDAR 3D perception model robustness under simultaneous domain shifts and label-space evolution in autonomous driving scenarios.

Ax Laura Dietz, Bryan Li, Gabrielle Liu, Jia-Huei Ju, Eugene Yang, Dawn Lawrie, William Walden, James Mayfield 3/30/2026

Incorporating Q&A Nuggets into Retrieval-Augmented Generation

Crucible system augments RAG with Q&A nuggets from documents, preserving citation provenance and improving extraction, selection, and report generation.

Ax Laura Dietz, Bryan Li, Eugene Yang, Dawn Lawrie, William Walden, James Mayfield 3/30/2026

Insider Knowledge: How Much Can RAG Systems Gain from Evaluation Secrets?

Study examining risks of RAG system evaluation and optimization using LLM judges, revealing circularity issues in nugget-based evaluation approaches.

Ax Donghee Lee, Rui Cai, Zhe Zhao 3/30/2026

CARPE: Context-Aware Image Representation Prioritization via Ensemble for Large Vision-Language Models

CARPE method improving vision-centric capabilities of vision-language models through context-aware image representation prioritization via ensemble approach.

Ax Kei Saito 3/30/2026

NRR-Phi: Text-to-State Mapping for Ambiguity Preservation in LLM Inference

Framework addressing LLM's tendency to collapse ambiguous inputs prematurely by mapping text to non-collapsing state spaces for better dialogue reasoning.

Ax Bhada Yun, Renn Su, April Yi Wang 3/30/2026

AI and My Values: User Perceptions of LLMs' Ability to Extract, Embody, and Explain Human Values from Casual Conversations

Study introducing VAPT toolkit to evaluate how LLMs extract, embody, and explain human values from conversations through user perception research.

Ax Weiyu Sun, Liangliang Chen, Yongnuo Cai, Huiru Xie, Yi Zeng, Ying Zhang 3/30/2026

EDU-CIRCUIT-HW: Evaluating Multimodal Large Language Models on Real-World University-Level STEM Student Handwritten Solutions

Benchmark for evaluating multimodal LLMs on handwritten STEM student solutions with mathematical formulas and diagrams, addressing authentic domain-specific evaluation gaps.

Ax Nisharg Nargund, Priyesh Shukla 3/30/2026

TernaryLM: Memory-Efficient Language Modeling via Native 1.5-Bit Quantization with Adaptive Layer-wise Scaling

TernaryLM: Language model trained natively with 1.5-bit quantization achieving memory-efficient deployment on edge devices while maintaining language modeling capability.

Ax Xiangbo Gao, Renjie Li, Xinghao Chen, Yuheng Wu, Suofei Feng, Qing Yin, Zhengzhong Tu 3/30/2026

PISCO: Precise Video Instance Insertion with Sparse Control

Video generation model for precise instance insertion with sparse control in filmmaking applications, moving beyond prompt-engineering toward controllable generation.

Ax Jared Zhu, Minhao Hu, Junde Wu 3/30/2026

SWE Context Bench: A Benchmark for Context Learning in Coding

Benchmark evaluating LLM-based coding agents on their ability to learn from context and reuse experience across related software engineering tasks in repositories.

Ax Nicholas Caputo 3/30/2026

Administrative Law's Fourth Settlement: AI and the Capability-Accountability Trap

Administrative law analysis of how government agencies balance technological capability with democratic oversight and accountability mechanisms.

Ax Manfred M. Fischer, Joshua Pitts 3/30/2026

The Effective Depth Paradox: Evaluating the Relationship between Architectural Topology and Trainability in Deep CNNs

Comparative study of CNN architectures (VGG, ResNet, GoogLeNet) analyzing relationship between depth and trainability in image recognition.

Ax Aditya Kumar Singh, Hitesh Kandala, Pratik Prabhanjan Brahma, Zicheng Liu, Emad Barsoum 3/30/2026

DUET-VLM: Dual stage Unified Efficient Token reduction for VLM Training and Inference

DUET-VLM: dual-stage token reduction framework for vision-language models reducing computational cost while maintaining accuracy during training and inference.

Ax Injun Baek, Yearim Kim, Nojun Kwak 3/30/2026

PedaCo-Gen: Scaffolding Pedagogical Agency in Human-AI Collaborative Video Authoring

PedaCo-Gen: pedagogically-informed human-AI system for collaborative instructional video generation using Cognitive Theory of Multimedia Learning.

Ax Shrestha Datta, Hongfu Liu, Anshuman Chhabra 3/30/2026

Golden Layers and Where to Find Them: Improved Knowledge Editing for Large Language Models Via Layer Gradient Analysis

Layer gradient analysis method for identifying optimal layers in LLMs for knowledge editing while preserving model behavior on unrelated inputs.

Ax Oliver Hoidn, Aashwin Mishra, Steven Henke, Albert Vong, Matthew Seaberg 3/30/2026

Towards single-shot coherent imaging via overlap-free ptychography

Extension of ptychographic imaging to overlap-free single-shot coherent diffractive imaging using physics-informed neural networks.

Ax Andrew Tremante, Yang He, Rocky Klopfenstein, Yuepeng Wang, Nina Narodytska, Haoze Wu 3/30/2026

SpotIt+: Verification-based Text-to-SQL Evaluation with Database Constraints

SpotIt+: open-source verification tool for Text-to-SQL evaluation using bounded equivalence checking and constraint-mining for practical query discrepancies.

Ax Ngoc-Son Nguyen, Thanh V. T. Tran, Jeongsoo Choi, Hieu-Nghia Huynh-Nguyen, Truong-Son Hy, Van Nguyen 3/30/2026

DiFlowDubber: Discrete Flow Matching for Automated Video Dubbing via Cross-Modal Alignment and Synchronization

DiFlowDubber: two-stage approach for automated video dubbing using discrete flow matching for expressive prosody and precise audio-visual synchronization.

Ax Xiangbo Gao, Mingyang Wu, Siyuan Yang, Jiongze Yu, Pardis Taghavi, Fangzhou Lin, Zhengzhong Tu 3/30/2026

The Pulse of Motion: Measuring Physical Frame Rate from Visual Dynamics

Method for measuring physical frame rate from visual dynamics in generative video models to improve temporal consistency.

Ax Zhaohui Geoffrey Wang 3/30/2026

AgentTrace: Causal Graph Tracing for Root Cause Analysis in Deployed Multi-Agent Systems

AgentTrace: lightweight framework for post-hoc root cause analysis in deployed multi-agent systems using causal graph tracing from execution logs.

Ax Yitong Zhang, Chengze Li, Ruize Chen, Guowei Yang, Xiaoran Jia, Yijie Ren, Jia Li 3/30/2026

To See is Not to Master: Teaching LLMs to Use Private Libraries for Code Generation

Study showing LLMs struggle with private library code generation despite API documentation; proposes teaching methods for private-library-oriented code generation.

Ax Redwan Sony, Anil K Jain, Arun Ross 3/30/2026

MLLM-based Textual Explanations for Face Comparison

Analysis of multimodal LLMs generating natural language explanations for face verification decisions on unconstrained images.

Ax Zenan Li, Ziran Yang, Deyuan He, Haoyu Zhao, Andrew Zhao, Shange Tang, Kaiyu Yang, Aarti Gupta, Zhendong Su, Chi Jin 3/30/2026

Goedel-Code-Prover: Hierarchical Proof Search for Open State-of-the-Art Code Verification

Goedel-Code-Prover: hierarchical proof search framework for automated code verification in Lean 4 using LLMs to decompose complex verification goals.

Ax Chien-Ping Lu 3/30/2026

Modernizing Amdahl's Law: How AI Scaling Laws Shape Computer Architecture

Analysis of how AI scaling laws reshape classical Amdahl's Law for modern heterogeneous computer architectures with specialized accelerators and tensor datapaths.

Ax Shuai Wang, Yinan Yu 3/30/2026

KG-Hopper: Empowering Compact Open LLMs with Knowledge Graph Reasoning via Reinforcement Learning

KG-Hopper: reinforcement learning framework enabling compact open-source LLMs to perform knowledge graph reasoning for multi-hop KBQA tasks.