# Athena AI News — Top 50 (Today Only) > Generated 2026-07-11 15:01 UTC · fresh filter: ingested 2026-07-11 (UTC) · ranked by virality + 18h decay > Source: oracle.db · 50 items shown of 80 fresh today | # | Title | Source | Score | Age | Link | |---|-------|--------|-------|-----|------| | 1 | Apple sues OpenAI, accuses ex-employees of stealing trade secrets | hackernews | 0.53 | 2h | [link](https://9to5mac.com/2026/07/10/apple-sues-openai-trade-secret-theft/) | | 2 | GPT-5.6 | hackernews | 0.51 | 2h | [link](https://openai.com/index/gpt-5-6/) | | 3 | texts-to-transformer: Train a tiny Transformer from scratch on your iMessage history, entirely on your Mac. | github | 0.47 | 2h | [link](https://github.com/Doriandarko/texts-to-transformer) | | 4 | GPT-5.6 Sol Ultra produces proof of the Cycle Double Cover Conjecture [pdf] | hackernews | 0.45 | 2h | [link](https://cdn.openai.com/pdf/04d1d1e4-bc75-476a-97cf-49055cd98d31/cdc_proof.pdf) | | 5 | Cognitive-Core-Skills: A universal, industry-neutral taxonomy of cognitive core skills (perception, memory, reasoning, plan | github | 0.43 | 2h | [link](https://github.com/eli-labz/Cognitive-Core-Skills) | | 6 | photoshop-ai-smart-enhance: AI-powered image enhancement extension for Adobe Photoshop CC 2024+. Intelligent exposure correction | github | 0.42 | 2h | [link](https://github.com/FuelMagistrateLead/photoshop-ai-smart-enhance) | | 7 | aipath: Interactive AI General Education Course — 30 Lessons, Zero Math | github | 0.39 | 2h | [link](https://github.com/buynao/aipath) | | 8 | AI-generated videos to maximally drive a target brain region | hackernews | 0.38 | 2h | [link](https://nevo-project.epfl.ch/) | | 9 | How the terrorist group Boko Haram uses frontier AI | hackernews | 0.38 | 2h | [link](https://casp.ac/reports/ai-enabled-terrorism) | | 10 | AI 2040: Plan A | hackernews | 0.38 | 2h | [link](https://ai-2040.com/) | | 11 | ChatGPT Work | hackernews | 0.37 | 2h | [link](https://openai.com/index/chatgpt-for-your-most-ambitious-work/) | | 12 | AI content is everywhere on social media, especially LinkedIn | hackernews | 0.36 | 2h | [link](https://www.pangram.com/blog/ai-in-your-feed) | | 13 | Building a real-time AI tutor for 5-year-olds | hackernews | 0.35 | 2h | [link](https://www.ello.com/blog/teaching-a-child-in-1000-ms) | | 14 | A font that humans can read but AI cannot | hackernews | 0.34 | 2h | [link](https://www.mixfont.com/ghost-font) | | 15 | GPT-5.6, Grok 4.5, Claude, and Muse Spark build the same 4 apps | hackernews | 0.34 | 2h | [link](https://www.tryai.dev/blog/gpt-5.6-build-off-12-models) | | 16 | reality-engine: Top Dynamic AI World Simulation & Storytelling Tools 2026 | github | 0.33 | 2h | [link](https://github.com/grandgaming9321-prog/reality-engine) | | 17 | manuscript-phoneme-decipher: Voynich Manuscript Decoded: Elu-Sinhala Phonetic Transcription & Vocabulary Toolkit 2026 | github | 0.33 | 2h | [link](https://github.com/okesipoke/manuscript-phoneme-decipher) | | 18 | ai-image-clean-eraser: AI-Powered Text Remover 2026: Auto-Detect & Manual Precision with HD Quality | github | 0.33 | 2h | [link](https://github.com/Sujal-142/ai-image-clean-eraser) | | 19 | churn-triad-insights: LLM-Powered Churn Risk Analyzer for Scalable 2026 Decision Support | github | 0.33 | 2h | [link](https://github.com/pravin6688/churn-triad-insights) | | 20 | swarm-foraging-qlearn: Q-Learning Swarm Foraging 2026: Multi-Agent RL in Dynamic Grid Environments | github | 0.33 | 2h | [link](https://github.com/jaimasih05-commits/swarm-foraging-qlearn) | | 21 | Paradigm-Survival-Arena: Top 6 AI Paradigms Fighting for Survival in 2026 | github | 0.33 | 2h | [link](https://github.com/aminekago-web/Paradigm-Survival-Arena) | | 22 | cortex-sentinel-trading-nexus: Self-Tuning Multi-Agent AI Trading System 2026: 8-Source Signal Fusion & Kronos Model | github | 0.33 | 2h | [link](https://github.com/reunios2024/cortex-sentinel-trading-nexus) | | 23 | magic-eraser-studio: AI Object Remover 2026 – Erase Distractions & Keep HD Quality | github | 0.33 | 2h | [link](https://github.com/onlyoneshakibul/magic-eraser-studio) | | 24 | ShipGenAI: 🚀 50 production-ready Generative AI SaaS apps — brand them, ship them, keep 100% of the revenue. Str | github | 0.32 | 2h | [link](https://github.com/benlamiro/ShipGenAI) | | 25 | ESEILANE: High-performance Knowledge Graph engine for AI, LLMs, and GraphRAG — built for the next generation o | github | 0.32 | 2h | [link](https://github.com/Aliu-AiRobot/ESEILANE) | | 26 | Hello-Agents: 🤖 Building AI Agent Systems from Scratch — A comprehensive, practical tutorial from fundamentals to | github | 0.31 | 2h | [link](https://github.com/Reyzowter/Hello-Agents) | | 27 | Apple sues OpenAI, accusing it of stealing company secrets | hackernews | 0.30 | 2h | [link](https://www.nytimes.com/2026/07/10/technology/apple-openai-lawsuit.html) | | 28 | ESEILANE: High-performance Knowledge Graph engine for AI, LLMs, and GraphRAG — built for the next generation o | github | 0.28 | 2h | [link](https://github.com/Simpl3x3/ESEILANE) | | 29 | Agent-Loop-Skills: Loop until it's better — drop-in agentic loops (autoresearch, scientific writing, data analysis, cod | github | 0.28 | 2h | [link](https://github.com/gaasher/Agent-Loop-Skills) | | 30 | autoguardrails: Alignment-research scaffold (autoresearch-style) for LLM guardrails: search over a single policy.md | github | 0.28 | 2h | [link](https://github.com/SantanderAI/autoguardrails) | | 31 | Anti-Autoresearch: Don't trust an autoresearch paper at face value. Reviewer-side integrity forensics (self-consistency | github | 0.28 | 2h | [link](https://github.com/wanshuiyin/Anti-Autoresearch) | | 32 | Ben Bernanke Joins Anthropic Oversight Trust | hackernews | 0.28 | 2h | [link](https://www.anthropic.com/news/ben-bernanke) | | 33 | SimPolitics: America’s quest to solve politics with computers | hackernews | 0.27 | 2h | [link](https://mitpress.mit.edu/9780262053198/simpolitics/) | | 34 | FerryAI: Native AI inference for PHP 8.3+ - run ONNX, GGUF (llama.cpp) and RubixML models directly in your PH | github | 0.27 | 2h | [link](https://github.com/MADEVAL/FerryAI) | | 35 | Show HN: FableCut – A browser video editor AI agents can drive (zero deps) | hackernews | 0.27 | 2h | [link](https://github.com/ronak-create/FableCut) | | 36 | Hands-On with the AMD Ryzen AI Halo | hackernews | 0.26 | 2h | [link](https://www.microcenter.com/site/mc-news/article/amd-ryzen-ai-halo-review.aspx) | | 37 | Show HN: Reverse-engineering web apps into agent tools | hackernews | 0.26 | 2h | [link](https://news.ycombinator.com/item/48847834) | | 38 | How version control will evolve for the agent boom | hackernews | 0.25 | 2h | [link](https://entire.io/blog/how-version-control-will-evolve-for-the-agent-boom) | | 39 | Show HN: Reviving my 2001 college band with AI | hackernews | 0.25 | 2h | [link](https://www.fadingmaize.com) | | 40 | The next era of AI is about infrastructure, not just models | hackernews | 0.23 | 2h | [link](https://blog.mozilla.ai/the-control-layer-why-the-next-era-of-ai-is-about-infrastructure-not-just-models/) | | 41 | UniClawBench: A Universal Benchmark for Proactive Agents on Real-World Tasks | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08768v1) | | 42 | OpenCoF: Learning to Reason Through Video Generation | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08763v1) | | 43 | Ideas Have Genomes: Benchmarking Scientific Lineage Reasoning and Lineage-Grounded Idea Generation | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08758v1) | | 44 | Score Accuracy Along the Forward Diffusion Does Not Certify Numerical Stability in Diffusion Sampling | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08757v1) | | 45 | MulTTiPop: A Multitrack Transcription Dataset for Pop Music | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08756v1) | | 46 | SLORR: Simple and Efficient In-Training Low-Rank Regularization | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08754v1) | | 47 | Using AI-based Learning Assistants in Higher Education: A Large-Scale Descriptive Analysis | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08748v1) | | 48 | Dimensionality Reduction Meets Network Science: Sensemaking on UMAP's kNN Graph | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08746v1) | | 49 | AUTOPILOT VQA: Benchmarking Vision-Language Models for Incident-Centric Dashcam Understanding | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08745v1) | | 50 | ARDY: Autoregressive Diffusion with Hybrid Representation for Interactive Human Motion Generation | arxiv | 0.00 | 2h | [link](https://arxiv.org/abs/2607.08741v1) | ## One-liners 3. **texts-to-transformer: Train a tiny Transformer from scratch ** — Train a tiny language model from scratch on your iMessage history, entirely on your Mac. 5. **Cognitive-Core-Skills: A universal, industry-neutral taxonom** — Cognitive core skills are the mental operating capabilities an LLM or AI Agent needs to move from chat response to useful digital co-worker. 6. **photoshop-ai-smart-enhance: AI-powered image enhancement ext** — photoshop-ai-smart-enhance is a machine learning extension that automates image quality improvements. 7. **aipath: Interactive AI General Education Course — 30 Lessons** — ↑ The homepage hero (Lesson 15 · the next-token game): an LLM guesses one token at a time — turn the temperature and watch its top-5 candidates reshape, from fo 16. **reality-engine: Top Dynamic AI World Simulation & Storytelli** — Welcome to **Chronos Engine** — a groundbreaking temporal simulation platform that empowers researchers, storytellers, game developers, and futurists to constru 17. **manuscript-phoneme-decipher: Voynich Manuscript Decoded: Elu** — What if a manuscript wasn't written in a lost language, but in a forgotten way of hearing. 18. **ai-image-clean-eraser: AI-Powered Text Remover 2026: Auto-De** — In a digital ecosystem where document fraud costs organizations over **$1. 19. **churn-triad-insights: LLM-Powered Churn Risk Analyzer for Sc** — Every day, thousands of customers quietly signal their intent to leave. 20. **swarm-foraging-qlearn: Q-Learning Swarm Foraging 2026: Multi** — Embark on a journey into emergent intelligence, where autonomous agents learn to collaborate, compete, and coexist in a living, breathing digital ecosystem. 21. **Paradigm-Survival-Arena: Top 6 AI Paradigms Fighting for Sur** — Unlike traditional ML benchmarks that test accuracy on static datasets, Synaptic Colosseum evaluates models on *adaptive fitness*: the ability to learn from spa 41. **UniClawBench: A Universal Benchmark for Proactive Agents on ** — we introduce UniClawBench, the first capability-driven benchmark designed to evaluate proactive agents in dynamic, real-world settings. 42. **OpenCoF: Learning to Reason Through Video Generation** — Reasoning has become a core capability for large models, especially when reliable decisions require understanding logical consequences 43. **Ideas Have Genomes: Benchmarking Scientific Lineage Reasonin** — We present IdeaGene-Bench (IG-Bench), a benchmark for scientific lineage reasoning and lineage-grounded idea generation. 44. **Score Accuracy Along the Forward Diffusion Does Not Certify ** — We show that small forward-marginal error does not guarantee numerical stability. 45. **MulTTiPop: A Multitrack Transcription Dataset for Pop Music** — We present MulTTiPop, a dataset of pop music segments and their associated multitrack MIDI recordings for the evaluation of automatic music transcription models 46. **SLORR: Simple and Efficient In-Training Low-Rank Regularizat** — Low-rank factorization is widely used to compress neural networks, but modern models are often not naturally amenable to aggressive factorization without signif 47. **Using AI-based Learning Assistants in Higher Education: A La** — we present a large-scale descriptive analysis of the use of an AI-based learning assistant (Syntea) in higher education. 48. **Dimensionality Reduction Meets Network Science: Sensemaking ** — we show that these graph-based analyses are not only practical but also competitive with or complementary to purpose-built methods (e. 49. **AUTOPILOT VQA: Benchmarking Vision-Language Models for Incid** — we present AUTOPILOT-VQA, an incident-centric visual question answering benchmark for dashcam video understanding. 50. **ARDY: Autoregressive Diffusion with Hybrid Representation fo** — Generating realistic 3D human motions in real-time within interactive applications is key for animation, simulation, and humanoid robotics --- ## Raw data (JSON) ```json [ { "title": "Apple sues OpenAI, accuses ex-employees of stealing trade secrets", "url": "https://9to5mac.com/2026/07/10/apple-sues-openai-trade-secret-theft/", "source": "hackernews", "score": 0.5253, "age": 2.0, "one": "" }, { "title": "GPT-5.6", "url": "https://openai.com/index/gpt-5-6/", "source": "hackernews", "score": 0.509, "age": 2.0, "one": "" }, { "title": "texts-to-transformer: Train a tiny Transformer from scratch on your iMessage history, entirely on your Mac.", "url": "https://github.com/Doriandarko/texts-to-transformer", "source": "github", "score": 0.4688, "age": 2.0, "one": "Train a tiny language model from scratch on your iMessage history, entirely on your Mac." }, { "title": "GPT-5.6 Sol Ultra produces proof of the Cycle Double Cover Conjecture [pdf]", "url": "https://cdn.openai.com/pdf/04d1d1e4-bc75-476a-97cf-49055cd98d31/cdc_proof.pdf", "source": "hackernews", "score": 0.448, "age": 2.0, "one": "" }, { "title": "Cognitive-Core-Skills: A universal, industry-neutral taxonomy of cognitive core skills (perception, memory, reasoning, plan", "url": "https://github.com/eli-labz/Cognitive-Core-Skills", "source": "github", "score": 0.4256, "age": 2.0, "one": "Cognitive core skills are the mental operating capabilities an LLM or AI Agent needs to move from chat response to useful digital co-worker." }, { "title": "photoshop-ai-smart-enhance: AI-powered image enhancement extension for Adobe Photoshop CC 2024+. Intelligent exposure correction", "url": "https://github.com/FuelMagistrateLead/photoshop-ai-smart-enhance", "source": "github", "score": 0.4197, "age": 2.0, "one": "photoshop-ai-smart-enhance is a machine learning extension that automates image quality improvements." }, { "title": "aipath: Interactive AI General Education Course — 30 Lessons, Zero Math", "url": "https://github.com/buynao/aipath", "source": "github", "score": 0.3918, "age": 2.0, "one": "↑ The homepage hero (Lesson 15 · the next-token game): an LLM guesses one token at a time — turn the temperature and watch its top-5 candidates reshape, from focused to “wild." }, { "title": "AI-generated videos to maximally drive a target brain region", "url": "https://nevo-project.epfl.ch/", "source": "hackernews", "score": 0.3827, "age": 2.0, "one": "" }, { "title": "How the terrorist group Boko Haram uses frontier AI", "url": "https://casp.ac/reports/ai-enabled-terrorism", "source": "hackernews", "score": 0.3793, "age": 2.0, "one": "" }, { "title": "AI 2040: Plan A", "url": "https://ai-2040.com/", "source": "hackernews", "score": 0.3785, "age": 2.0, "one": "" }, { "title": "ChatGPT Work", "url": "https://openai.com/index/chatgpt-for-your-most-ambitious-work/", "source": "hackernews", "score": 0.3745, "age": 2.0, "one": "" }, { "title": "AI content is everywhere on social media, especially LinkedIn", "url": "https://www.pangram.com/blog/ai-in-your-feed", "source": "hackernews", "score": 0.3553, "age": 2.0, "one": "" }, { "title": "Building a real-time AI tutor for 5-year-olds", "url": "https://www.ello.com/blog/teaching-a-child-in-1000-ms", "source": "hackernews", "score": 0.3519, "age": 2.0, "one": "" }, { "title": "A font that humans can read but AI cannot", "url": "https://www.mixfont.com/ghost-font", "source": "hackernews", "score": 0.3433, "age": 2.0, "one": "" }, { "title": "GPT-5.6, Grok 4.5, Claude, and Muse Spark build the same 4 apps", "url": "https://www.tryai.dev/blog/gpt-5.6-build-off-12-models", "source": "hackernews", "score": 0.343, "age": 2.0, "one": "" }, { "title": "reality-engine: Top Dynamic AI World Simulation & Storytelling Tools 2026", "url": "https://github.com/grandgaming9321-prog/reality-engine", "source": "github", "score": 0.3342, "age": 2.0, "one": "Welcome to **Chronos Engine** — a groundbreaking temporal simulation platform that empowers researchers, storytellers, game developers, and futurists to construct, explore, and manipulate dynamic time" }, { "title": "manuscript-phoneme-decipher: Voynich Manuscript Decoded: Elu-Sinhala Phonetic Transcription & Vocabulary Toolkit 2026", "url": "https://github.com/okesipoke/manuscript-phoneme-decipher", "source": "github", "score": 0.3342, "age": 2.0, "one": "What if a manuscript wasn't written in a lost language, but in a forgotten way of hearing." }, { "title": "ai-image-clean-eraser: AI-Powered Text Remover 2026: Auto-Detect & Manual Precision with HD Quality", "url": "https://github.com/Sujal-142/ai-image-clean-eraser", "source": "github", "score": 0.3342, "age": 2.0, "one": "In a digital ecosystem where document fraud costs organizations over **$1." }, { "title": "churn-triad-insights: LLM-Powered Churn Risk Analyzer for Scalable 2026 Decision Support", "url": "https://github.com/pravin6688/churn-triad-insights", "source": "github", "score": 0.3335, "age": 2.0, "one": "Every day, thousands of customers quietly signal their intent to leave." }, { "title": "swarm-foraging-qlearn: Q-Learning Swarm Foraging 2026: Multi-Agent RL in Dynamic Grid Environments", "url": "https://github.com/jaimasih05-commits/swarm-foraging-qlearn", "source": "github", "score": 0.3335, "age": 2.0, "one": "Embark on a journey into emergent intelligence, where autonomous agents learn to collaborate, compete, and coexist in a living, breathing digital ecosystem." }, { "title": "Paradigm-Survival-Arena: Top 6 AI Paradigms Fighting for Survival in 2026", "url": "https://github.com/aminekago-web/Paradigm-Survival-Arena", "source": "github", "score": 0.3335, "age": 2.0, "one": "Unlike traditional ML benchmarks that test accuracy on static datasets, Synaptic Colosseum evaluates models on *adaptive fitness*: the ability to learn from sparse rewards, generalize from limited exa" }, { "title": "cortex-sentinel-trading-nexus: Self-Tuning Multi-Agent AI Trading System 2026: 8-Source Signal Fusion & Kronos Model", "url": "https://github.com/reunios2024/cortex-sentinel-trading-nexus", "source": "github", "score": 0.3335, "age": 2.0, "one": "" }, { "title": "magic-eraser-studio: AI Object Remover 2026 – Erase Distractions & Keep HD Quality", "url": "https://github.com/onlyoneshakibul/magic-eraser-studio", "source": "github", "score": 0.3335, "age": 2.0, "one": "" }, { "title": "ShipGenAI: 🚀 50 production-ready Generative AI SaaS apps — brand them, ship them, keep 100% of the revenue. Str", "url": "https://github.com/benlamiro/ShipGenAI", "source": "github", "score": 0.324, "age": 2.0, "one": "" }, { "title": "ESEILANE: High-performance Knowledge Graph engine for AI, LLMs, and GraphRAG — built for the next generation o", "url": "https://github.com/Aliu-AiRobot/ESEILANE", "source": "github", "score": 0.3238, "age": 2.0, "one": "" }, { "title": "Hello-Agents: 🤖 Building AI Agent Systems from Scratch — A comprehensive, practical tutorial from fundamentals to ", "url": "https://github.com/Reyzowter/Hello-Agents", "source": "github", "score": 0.3146, "age": 2.0, "one": "" }, { "title": "Apple sues OpenAI, accusing it of stealing company secrets", "url": "https://www.nytimes.com/2026/07/10/technology/apple-openai-lawsuit.html", "source": "hackernews", "score": 0.3006, "age": 2.0, "one": "" }, { "title": "ESEILANE: High-performance Knowledge Graph engine for AI, LLMs, and GraphRAG — built for the next generation o", "url": "https://github.com/Simpl3x3/ESEILANE", "source": "github", "score": 0.2844, "age": 2.0, "one": "" }, { "title": "Agent-Loop-Skills: Loop until it's better — drop-in agentic loops (autoresearch, scientific writing, data analysis, cod", "url": "https://github.com/gaasher/Agent-Loop-Skills", "source": "github", "score": 0.2841, "age": 2.0, "one": "" }, { "title": "autoguardrails: Alignment-research scaffold (autoresearch-style) for LLM guardrails: search over a single policy.md ", "url": "https://github.com/SantanderAI/autoguardrails", "source": "github", "score": 0.284, "age": 2.0, "one": "" }, { "title": "Anti-Autoresearch: Don't trust an autoresearch paper at face value. Reviewer-side integrity forensics (self-consistency", "url": "https://github.com/wanshuiyin/Anti-Autoresearch", "source": "github", "score": 0.2834, "age": 2.0, "one": "" }, { "title": "Ben Bernanke Joins Anthropic Oversight Trust", "url": "https://www.anthropic.com/news/ben-bernanke", "source": "hackernews", "score": 0.2826, "age": 2.0, "one": "" }, { "title": "SimPolitics: America’s quest to solve politics with computers", "url": "https://mitpress.mit.edu/9780262053198/simpolitics/", "source": "hackernews", "score": 0.273, "age": 2.0, "one": "" }, { "title": "FerryAI: Native AI inference for PHP 8.3+ - run ONNX, GGUF (llama.cpp) and RubixML models directly in your PH", "url": "https://github.com/MADEVAL/FerryAI", "source": "github", "score": 0.2726, "age": 2.0, "one": "" }, { "title": "Show HN: FableCut – A browser video editor AI agents can drive (zero deps)", "url": "https://github.com/ronak-create/FableCut", "source": "hackernews", "score": 0.2723, "age": 2.0, "one": "" }, { "title": "Hands-On with the AMD Ryzen AI Halo", "url": "https://www.microcenter.com/site/mc-news/article/amd-ryzen-ai-halo-review.aspx", "source": "hackernews", "score": 0.2582, "age": 2.0, "one": "" }, { "title": "Show HN: Reverse-engineering web apps into agent tools", "url": "https://news.ycombinator.com/item/48847834", "source": "hackernews", "score": 0.2561, "age": 2.0, "one": "" }, { "title": "How version control will evolve for the agent boom", "url": "https://entire.io/blog/how-version-control-will-evolve-for-the-agent-boom", "source": "hackernews", "score": 0.2506, "age": 2.0, "one": "" }, { "title": "Show HN: Reviving my 2001 college band with AI", "url": "https://www.fadingmaize.com", "source": "hackernews", "score": 0.2497, "age": 2.0, "one": "" }, { "title": "The next era of AI is about infrastructure, not just models", "url": "https://blog.mozilla.ai/the-control-layer-why-the-next-era-of-ai-is-about-infrastructure-not-just-models/", "source": "hackernews", "score": 0.2311, "age": 2.0, "one": "" }, { "title": "UniClawBench: A Universal Benchmark for Proactive Agents on Real-World Tasks", "url": "https://arxiv.org/abs/2607.08768v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "we introduce UniClawBench, the first capability-driven benchmark designed to evaluate proactive agents in dynamic, real-world settings." }, { "title": "OpenCoF: Learning to Reason Through Video Generation", "url": "https://arxiv.org/abs/2607.08763v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "Reasoning has become a core capability for large models, especially when reliable decisions require understanding logical consequences" }, { "title": "Ideas Have Genomes: Benchmarking Scientific Lineage Reasoning and Lineage-Grounded Idea Generation", "url": "https://arxiv.org/abs/2607.08758v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "We present IdeaGene-Bench (IG-Bench), a benchmark for scientific lineage reasoning and lineage-grounded idea generation." }, { "title": "Score Accuracy Along the Forward Diffusion Does Not Certify Numerical Stability in Diffusion Sampling", "url": "https://arxiv.org/abs/2607.08757v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "We show that small forward-marginal error does not guarantee numerical stability." }, { "title": "MulTTiPop: A Multitrack Transcription Dataset for Pop Music", "url": "https://arxiv.org/abs/2607.08756v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "We present MulTTiPop, a dataset of pop music segments and their associated multitrack MIDI recordings for the evaluation of automatic music transcription models." }, { "title": "SLORR: Simple and Efficient In-Training Low-Rank Regularization", "url": "https://arxiv.org/abs/2607.08754v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "Low-rank factorization is widely used to compress neural networks, but modern models are often not naturally amenable to aggressive factorization without significant accuracy loss" }, { "title": "Using AI-based Learning Assistants in Higher Education: A Large-Scale Descriptive Analysis", "url": "https://arxiv.org/abs/2607.08748v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "we present a large-scale descriptive analysis of the use of an AI-based learning assistant (Syntea) in higher education." }, { "title": "Dimensionality Reduction Meets Network Science: Sensemaking on UMAP's kNN Graph", "url": "https://arxiv.org/abs/2607.08746v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "we show that these graph-based analyses are not only practical but also competitive with or complementary to purpose-built methods (e." }, { "title": "AUTOPILOT VQA: Benchmarking Vision-Language Models for Incident-Centric Dashcam Understanding", "url": "https://arxiv.org/abs/2607.08745v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "we present AUTOPILOT-VQA, an incident-centric visual question answering benchmark for dashcam video understanding." }, { "title": "ARDY: Autoregressive Diffusion with Hybrid Representation for Interactive Human Motion Generation", "url": "https://arxiv.org/abs/2607.08741v1", "source": "arxiv", "score": 0.0, "age": 2.0, "one": "Generating realistic 3D human motions in real-time within interactive applications is key for animation, simulation, and humanoid robotics" } ] ```