Software Engineering Hacker News

We put a Redis server inside our runtime

An engineering deep dive into integrating a Redis server directly into a runtime environment.

Hardware/Chips Hacker News

GLM5.2 on AMD MI355X at 2626 tok/s/node at over 2x lower cost than Blackwell

A performance benchmark comparing GLM5.2 performance on AMD MI355X hardware against NVIDIA Blackwell, showing significant cost-effectiveness.

Open Source Hacker News

GitFut – Your GitHub stats turned into a World-Cup-style player card

GitFut – A tool that transforms GitHub statistics into player cards styled after a World-Cup player card format.

Other Hacker News

The Demoralization of the White-Collar Worker

A discussion on the psychological state and demoralization of modern white-collar workers.

Cybersecurity Hacker News

New serious vulnerabilities spiked around release of Claude Mythos Preview

Reports on a spike in serious security vulnerabilities coinciding with the release of the Claude Mythos Preview.

Other Hacker News

Nerdle Review

A review of Nerdle, a math-based wordle-like game.

AI/ML arXiv cs.AI

ADVENT: LLM-Driven Automatic Predicate Invention for ILP

Introduces ADVENT, an LLM-driven mechanism for automatic predicate invention in Inductive Logic Programming to enable cross-task knowledge reuse.

AI/ML arXiv cs.AI

VLAFlow: A Unified Training Framework for Vision-Language-Action Models via Co-training and Future Latent Alignment

Presents VLAFlow, a unified training framework for Vision-Language-Action models to compare different pre-training paradigms in robotics.

AI/ML arXiv cs.AI

MKGR: Multimodal Knowledge-Graph Representation Learning for Cold-Start Protein-Protein Interaction Prediction

Proposes MKGR, a multimodal representation framework using knowledge graphs and protein sequences for cold-start protein-protein interaction prediction.

AI/ML arXiv cs.AI

AgenticDataBench: A Comprehensive Benchmark for Data Agents

Introduces AgenticDataBench, a comprehensive benchmark for evaluating LLM-based data agents across diverse real-world data science workflows.

Cybersecurity arXiv cs.AI

Beyond Gradient-Based Attacks: Adversarial Robustness and Explainability Stability in Cybersecurity Classifiers

Analyzes adversarial robustness and the stability of SHAP-based explanations in cybersecurity classifiers, introducing the Explainability Stability Index (ESI).

AI/ML arXiv cs.AI

Model Merging as Probabilistic Inference in Fine-Tuning Parameter Space

Formulates model merging as probabilistic inference under a product-of-experts scenario to improve multi-task solution combining without further fine-tuning.

Cybersecurity arXiv cs.AI

Pmeta-TLA: Backdoor Attacks for Speech Classification Models via Meta-Learning with Timbre Leakage Attack

Introduces Pmeta-TLA, a meta-learning training mechanism for embedding multiple stealthy backdoor attacks in speech classification models via timbre leakage.

Other Hacker News

Africans Are Turning to Starlink

Discussion on the increasing adoption of Starlink internet services across various African nations.

Other Hacker News

Goodbye, Forever, Probably

A personal or philosophical reflection piece titled 'Goodbye, Forever, Probably'.

Software Engineering Hacker News

Applied Category Theory Course (2018)

A reference to a 2018 course on Applied Category Theory, a mathematical framework used in computer science.

Software Engineering Hacker News

Show HN: ContextCodeCache in Rust

Introduction of ContextCodeCache, a Rust-based tool designed for efficient code context caching.

Other Hacker News

ClawdMojis – A Clawd for Every Occasion

A project creating 'ClawdMojis', custom emojis based on a specific character or theme.

AI/ML arXiv cs.AI

Multi-Head Recurrent Memory Agents

Proposes Multi-Head Recurrent Memory (MHM) to solve reliability issues in long-context LLMs by partitioning memory into independent heads.

AI/ML arXiv cs.AI

IntentTune: Using user demand and personalization to resolve "unknown" query intents for e-commerce search

Introduces IntentTune, a framework that uses user behavior and population patterns to resolve ambiguous search queries in e-commerce.