1003 related articles

A PKU-Stanford trainer breaks down how Python surpasses Stata and R, how AI-driven Skills and Paper Workflow automate empirical research from data to LaTeX paper drafts.

When RL continuously optimizes models to please reward models, do soaring Elo scores truly represent capability gains? A deep dive into Reward Hacking in RLHF, Goodhart's Law in AI, and industry countermeasures.

What happens when AI agents are tasked with running a real company? This analysis examines agent performance, critical shortcomings, and practical enterprise deployment advice.

Decoding signals like "frontiermogging" to analyze upcoming AI frontier model leaps, Agent automation deployment, and developer ecosystem expansion trends.

Harvard and UIUC propose a third axis of pretraining, claiming 6.2x sample efficiency and 250x inference speedup. Deep analysis of this new paradigm's implications and key caveats.

Deep dive into Heretic uncensoring technology applied to Jamba2-Mini, Qwen3.5-9B, and 27B open-source models, exploring how refusal rates dropped from 97% to 4% and the safety debates involved.

How the internet's core architecture was accidentally built by engineers solving specific problems—from TCP/IP to search engines to AI data infrastructure—revealing bottom-up emergence patterns.

Google Earth introduces AI-generated satellite imagery, drastically lowering the barrier for faking satellite photos. This article analyzes the impact on journalism, environmental monitoring, and legal evidence.

Explore how graph engineering uses state machines and directed graph structures to constrain AI agent behavior, covering reflection, routing, human-in-the-loop, and parallel execution patterns.

Deep dive into how graph engineering uses state machines and directed graphs to constrain AI agent behavior, covering reflection, routing, human-in-the-loop, and parallel execution patterns.

AI aces reasoning tests but may reason incorrectly. This article analyzes fake reasoning behind correct answers in LLMs, covering data contamination, memory effects, and methods like process supervision and counterfactual testing.

A user burned over 14,000 Perplexity Computer credits building an AI Agent workflow with zero output. We break down the three fatal flaws and the Claude+OpenAI workaround.

Halo is a local real-time deepfake detection tool that identifies AI-synthesized faces during Zoom, Teams, and Google Meet video calls to prevent face-swapping fraud.

Poth Labs models customer knowledge as a dynamic relationship network, using cross-source reasoning and adaptive surveys to help enterprises understand churn and feature adoption.

Deep dive into Google DeepMind's Gemini Robotics 2: its whole-body intelligence, dexterous manipulation, adaptive reasoning, and how multi-robot collaboration is advancing embodied AI from lab to reality.

Deep dive into Google DeepMind's Gemini Robotics 2: its three core capabilities of whole-body intelligence, dexterous manipulation, and adaptive reasoning, plus how multi-robot collaboration is pushing embodied AI from labs into the physical world.

A complete technical guide to automatic Tibetan-Chinese bilingual subtitle generation, covering Tibetan ASR (Whisper/wav2vec), machine translation (NLLB), timeline alignment, and subtitle export for low-resource language creators.

A reported 3-word jailbreak of Claude Opus 5 sparks debate. We analyze LLM jailbreak mechanics, alignment fragility, and defense-in-depth strategies for AI security.

How can PhD students avoid coding skill atrophy when using AI programming assistants? This article proposes a layered delegation strategy with actionable advice for researchers.

Two papers flagged for fake authors still received oral presentation slots at top conferences, exposing systemic peer review failures in the AI era.