Ax Diego Marcondes 19d ago

Random test functions, $H^{-1}$ norm equivalence, and stochastic variational physics-informed neural networks

Physics-informed neural networks using random test functions to compute H^-1 norm equivalence for solving PDEs.

Ax Wei Deng 19d ago

Weak-to-Strong Elicitation via Mismatched Wrong Drafts

Weak model drafts improve strong model GRPO training by injecting mathematically wrong domain-specific examples, outperforming standard on-policy RL.

Ax Tanmay Asthana, Aman Saksena, Divyansh Sahu 19d ago

Evaluating Deep Research Agents on Expert Consulting Work: A Benchmark with Verifiers, Rubrics, and Cognitive Traps

Benchmark evaluating deep research agents on expert consulting tasks with 70 SME-authored prompts embedding cognitive traps.

Ax Michel A. Youssef 19d ago

CALIBURN: Operationally Calibrated Streaming Intrusion Detection with Regime-Dependent Conformal Risk Control

Streaming intrusion detection system using conformal risk control for threshold-based alerting in cybersecurity operations.

Ax Augustin Chan 19d ago

Symbolic Reasoning Frameworks Trigger Memory-Mediated Ecosystem Dynamics in Multi-Agent LLM Systems

Multi-agent LLM systems with symbolic reasoning frameworks show emergent risk behavior changes through memory interactions and reflective prompts.

Ax Yash Vardhan Tomar, Dheeraj Peddireddy 19d ago

Graph Reinforcement Learning for Calibration-Aware Quantum Circuit Routing

Quantum circuit routing using reinforcement learning and calibration data from IBM Heron processors to optimize SWAP gates.

Ax Greg Ongie, Rahul Parhi 19d ago

Representation Costs in Data Science: Foundations and the Quasi-Banach Spaces of Deep Neural Networks

arXiv paper developing framework for analyzing representation costs in parametric data-fitting and deep neural networks via regularizers.

Ax Aman Pathak, Cheng Peng, Mengxian Lyu, Ziyi Chen, Reema Solan, Sankalp Talankar, Yasir Khan, Hiren Mehta, Aokun Chen, Yi Guo, Yonghui Wu 19d ago

Prompt, Plan, Extract: Zero-Shot Agentic LLMs Workflows for Lung Pathology Extraction from Clinical Narratives

arXiv paper on agentic LLM workflows for zero-shot information extraction from lung pathology reports using prompt-plan-extract approach.

Ax Zhan Shi, Bing He, Yisi Sang, Hanqing Lu, Benoit Dumoulin 19d ago

A-Evolve-Training: Autonomous Post-Training of a 30B Model

arXiv paper on A-Evolve-Training, autonomous post-training system for 30B model with no human-in-loop over multiple weeks.

Ax Junfeng Liu, Christopher T. Symons, Ranga Raju Vatsavai 19d ago

GRAG: Generic Response-Augmented Generation Framework for Personalized Conversational Systems

arXiv paper on GRAG framework for personalized conversational agents with separate treatment of content grounding and personalization.

Ax Yang Gao (Veyon Solutions) 19d ago

How Reliable Is Your Jailbreak Judge? Calibration and Adversarial Robustness of Automated ASR Scoring

arXiv paper evaluating reliability of automated jailbreak judges for LLM safety, comparing classifiers vs human evaluation.

Ax Ilia Kulikov, Chenxi Whitehouse, Tianhao Wu, Yixin Nie, Swarnadeep Saha, Eryk Helenowski, Weizhe Yuan, Olga Golovneva, Jack Lanchantin, Yoram Bachrach, Jakob Foerster, Xian Li, Han Fang, Sainbayar Sukhbaatar, Jason Weston 19d ago

Autodata: An agentic data scientist to create high quality synthetic data

arXiv paper introducing Autodata, an AI agent that acts as data scientist to generate synthetic training data via Agentic Self-Instruct.

HN gw5815 19d ago

Building domain-specific AI chatbots with Claude

Technical guide on building domain-specific expert chatbots using Claude with architectural patterns and two-model approach.

HN mzubairtahir 19d ago

Show HN: A GPU/VRAM filter for finding LLMs that will run on your hardware

GPU/VRAM filter tool for searching which LLMs will run on specific hardware at various quantization levels.

HN nsagent 19d ago

No Bailouts for Big Tech Billionaires: Policies for When the AI Bubble Bursts

Opinion essay comparing AI industry speculation to the 2008 housing bubble.

HN olalonde 19d ago

Ask HN: How has the ML engineer job evolved outside of LLM labs?

Discussion of how ML engineering roles have evolved as frontier LLMs replace custom-trained models in non-lab environments.

HN loa_observer 19d ago

Rura: Interactive TUI scratchpad for building shell pipelines

Terminal UI tool for building and testing shell pipelines interactively with live output and diff capabilities.

HN AaronO 19d ago

White House asks OpenAI to limit its next model release

White House requests OpenAI restrict GPT 5.6 release to government partners due to advanced capabilities.

HN snorbleck 19d ago

Guardrails for offensive AI agents and why instructions are never enough

Discusses guardrails and safety mechanisms for preventing AI agents from executing harmful instructions.

HN oogali 19d ago

Local-first AI coding assistant for IntelliJ-based IDEs (paid)

Local-first AI coding assistant for JetBrains IDEs with offline inference capabilities.

HN billybuckwheat 19d ago

AI is reshaping work. A new group aims to help people adapt and thrive

News article on AI workforce impact and policy initiatives to help workers adapt to automation.

HN Littice 19d ago

Show HN: Ludion – routing AI inference by observed WebGPU behavior

Ludion is a routing system for AI inference that optimizes WebGPU behavior and resource allocation.

HN noida 19d ago

Anthropic Accuses Alibaba of Largest AI Distillation Attack: 28.8M Fraudulent

Anthropic accuses Alibaba of orchestrating largest AI model distillation attack with 28.8M fraudulent exchanges.

HN LorenDB 19d ago

Linux Foundation Launches Akrites to Defend FOSS from AI-Enabled Exploits

Author bio and announcement page for Linux Foundation's Akrites project defending FOSS from AI exploits.

HN novashang 19d ago

DropItDown – Drop a file, get Markdown your AI agent can read (macOS)

DropItDown: macOS tool converting files to clean Markdown locally for AI agents without account or token burn.

HN Alien1Being 19d ago

How do you get good ideas for startups?

Paul Graham essay on generating startup ideas by questioning why ideation seems difficult.

HN binyu 19d ago

Thermodynamic Computing from Zero to One

Extropic announces breakthrough AI algorithms and hardware addressing energy efficiency as limiting factor for AI scaling.

HN ilreb 19d ago

CS2-10k: A Large-Scale Egocentric Counter-Strike 2 Dataset

CS2-10k dataset: large-scale egocentric Counter-Strike 2 video with synchronized action inputs for world model training.

HN CalvinSturm 19d ago

Show HN: FastPlay, a fast minimal Windows video player built in Rust

FastPlay: minimal Windows video player built in Rust optimized for fast file opening and responsive controls.

HN Aaron_Agassi 19d ago

Claude AI's link handling and fix

Link to article about Claude AI link handling with minimal content.

HN navs 19d ago

Hollywood-backed nonprofit launches machine-readable AI consent registry

RSL Media launches machine-readable AI consent registry allowing creators to signal approval, conditional use, or prohibition of AI on their work.

HN reboy 19d ago

Show HN: HoprLabs – a Python lab for prototyping AI math ideas

HoprLab: Python CLI toolkit for simulating AI training math, estimating model size, VRAM usage, training time, and config risks before actual model training.

HN arjunchint 19d ago

Show HN: DeepSeek Flash inverted the economics of agent products

DeepSeek Flash inverted AI agent economics: open-source model beats Sonnet on benchmarks, eliminates pricing subsidy dynamics favoring developers.

HN thewebguyd 19d ago

New macOS malware embeds fake errors to confuse AI analysis tools

macOS malware 'Gaslight' embeds fake errors and prompt injection strings to confuse AI-assisted malware analysis tools.

HN tanelpoder 19d ago

OpenAI will delay GPT-5.6 after Trump administration request

News placeholder about OpenAI delaying GPT-5.6 following Trump administration request with case-by-case government approval.

HN lekan_digital 19d ago

Export Claude.ai chats, artifacts, and visible thinking

Chrome extension for exporting Claude.ai conversations, artifacts, and visible thinking as organized ZIP files with local browser processing.

HN snyhlxde 19d ago

JetSpec Enables Up to 9.64x Lossless LLM Inference Speedup with Up to 1000TPS

JetSpec: LLM inference optimization achieving up to 9.64x speedup and 1000 tokens-per-second throughput using speculative decoding.

HN sebg 19d ago

General LLMs outperform specialized clinical AI tools on medical benchmarks

Study comparing general LLMs versus specialized clinical AI tools on medical benchmarks. No detailed findings or data provided.

HN htrp 19d ago

Trump administration asks OpenAI to stagger release of GPT5.6

Brief headline about Trump administration requesting OpenAI stagger GPT-5.6 release. No substantive details provided.

HN sona-coffee11 19d ago

Extract structured JSON from messy text with Telnyx AI Inference

Telnyx AI Inference service extracts structured JSON from unstructured text like support tickets and emails using configurable schemas.

HN sebg 19d ago

Scaling Laws, Carefully

Technical analysis of scaling laws in deep learning, their empirical foundations, and optimal compute allocation strategies.

HN sohkamyung 19d ago

Hidden Prompts in Manuscripts Exploit AI-Assisted Peer Review

News brief with no content: Hidden prompts in manuscripts exploit AI-assisted peer review.

HN sameersri2004 19d ago

Ask HN: Have you ever given your AI agent a phone number?

Discussion: Users sharing experiences giving AI agents phone numbers and real-world tool access for task automation.

HN I_am_tiberius 19d ago

The US Government has requested a slow staggered rollout of GPT-5.6

News headline snippet about US government requesting slow rollout of GPT-5.6. Minimal substantive content.

HN localghost3000 19d ago

Nb: Command line note‑taking, bookmarking, archiving, and knowledge base

Command-line note-taking and knowledge base application supporting multiple formats and encryption.

HN firetesterlab 19d ago

Pipestage

Open-source Python library for building async data pipelines with staged processing and bounded concurrency. Zero dependencies.

HN b-man 19d ago

All you need is PostgreSQL

Discussion on using PostgreSQL for various use cases instead of specialized tools. Software architecture opinion piece.

HN wonkyfruit 19d ago

Show HN: A Claude skill that prunes your AI's memory file, one diff at a time

Memory optimization tool for Claude that prunes bloated memory files using diffs to improve agent performance over time.

HN mooreds 19d ago

Why AI agents need three types of memory

Technical article on three types of memory required for effective AI agent operation and design.

HN skeledrew 19d ago

The Coming Divide: AI-Native or Left Behind

Headline about divide between AI-native companies and those left behind. No substantive content.