Curvature-aware MDL framework for layer-wise capacity allocation and pruning decisions in large language model optimization.
Studies sign identifiability of causal effects in continuous-time linear stochastic differential equations.
Goal-Driven Data Optimization framework for efficient multimodal instruction tuning of vision-language models, reducing training data needs.
Spatiotemporal Gaussian representation framework for time-resolved 3D MRI reconstruction in radiotherapy.
Tessera system for kernel-granularity GPU disaggregation to optimize heterogeneous GPU clusters for LLMs.
Analysis of validity challenges when using NLP embeddings as proxies for social science constructs.
ZAPS-DA reduces action jitter in continuous control RL policies without phase lag during deployment.
Loopzero benchmark for detecting collapse patterns in recursive systems using telemetry analysis.
SHARP method for learning long-range temporal patterns in streaming data without full sequence revisiting.
TNODEV toolbox for formal verification of neural ordinary differential equations in safety-critical systems.
InvestPhilBench evaluates LLMs on reconstructing and applying expert investment decision frameworks.
Theoretical analysis of grokking phenomenon using stochastic-geometric framework to explain delayed generalization.
Framework for governing generative AI risk in financial institutions, extending banking regulatory standards.
Local LLM used to develop fuzzy cognitive maps by extracting quantitative relationships from textual data.
Pelican-VLA 0.5 is a unified vision-language-action model for robotic manipulation without task-specific fine-tuning.
Research on motion control for quadruped robots using motion-capture data and humanoid control techniques.
Title only. Proposes CPU-only topological AI core alternative to transformer-based LLM runtime.
Kernel release announcement.
Open-source real-time analytics engine for operational data with LLM inference. Built on Apache DataFusion. v2.0 adds CDC replication.
Tutorial on building GitHub code review agent.
Opinion piece on LLM document handling risks. Limited technical analysis.
Self-hosted prompt bar enabling website self-editing using Claude Code.
Geopolitical claim about Anthropic tool security. Not technical analysis.
Tool mapping repositories to knowledge graphs for open-source matching.
Hardware specs list for AI computing. Minimal depth.
Tracker: Lightweight self-hosted document store for coding agents, written in Go with Postgres index and MCP server support.
Early literacy app using spaced repetition for teaching children to read.
Title only. Appears to be benchmark results, insufficient content.
Snippet about developer tool combining Go AST with LLM. Insufficient content for evaluation.
Title only. Benchmark for evaluating AI agents on real-world professional workflows.
Finterm.ai CLI tool enabling coding agents and LLMs direct access to financial data: stock prices, options, SEC filings for trading strategies.
Oh My Pi: Open-source coding agent with 40+ providers, 32 built-in tools, 55k lines of Rust core, available for macOS/Linux/Windows.
MCP server exposing Tunetank royalty-free music catalog to AI assistants like Claude and ChatGPT for audio/video content.
DRIVE framework for measuring engineering org health in AI era across delivery, reliability, initiatives, vigilance, efficiency metrics.
Title only. Vague tech complaint article.
The Onion satire video. Not substantive content.
Note that 95% of announced Nvidia Grace Blackwell GPUs remain undeployed.
LearnChess platform uses AI coach to teach chess through interactive lessons, puzzles, and plain-English explanations integrated with engine analysis.
Research on confidential AI execution using CPU/GPU trusted execution environments. Security-focused approach.
Open benchmark dataset for investor workflows and financial professional tasks. Largest dataset of its kind.
Framework for building full-stack apps with agents. Simplifies deployment from hobby to production. Agent-first architecture.
GPT-5.6 Sol model achieves 13.33% on ARC-AGI public benchmark, first to win game with multi-step reasoning.
PDF system card for GPT-5.6 model.
Open-source flight fare monitor detecting price errors before airlines correct them; 67-line detector.
Open-source CLI tool that auto-generates and updates markdown wikis from coding agent conversations. Self-hosted, integrates with existing agents.
Stub announcement of GPT-5.6 release.
Open-source distributed Firecracker microVM orchestrator written in Elixir with gRPC support.
Brief comparison of LLM capabilities for programming vs writing tasks. Limited technical depth.
Guide on using coding agents for reproducible data science and machine learning workflows.
HN discussion on sharing agent context across team members running local Claude Code agents.