AI/ML Link Batch — 2026-06-27 Lines 21-100
This page records ingestion of links 21-100 from links-no-label-aah-ai-gold-jjh-strix-2026-06-27.txt. Lines 1-20 were previously ingested in ai-ml-link-batch-2026-06-27. Line 58 was a duplicate of the Karpathy LLM wiki gist already captured in the first batch and was skipped.
Research-interest summary
Most interesting to a CS researcher:
- coding-agents-and-skills — agent scaffolding, skill-vs-doc evals, GEPA/DSPy-style text-parameter optimization, and agent execution environments.
- local-ai-hardware-and-inference — ROCm, Strix Halo, vLLM/Triton, llama.cpp, and local coding-agent benchmarks as systems evidence.
- open-models-and-gemma-4 — open-model progress plus Gemma 4 multi-token prediction as architecture/runtime co-design.
- llm-evaluation-and-belief-management — belief revision, forgetting, and statistics for LLM evals.
- world-models-and-video-intelligence and robotics-and-embodied-ai — world models, video intelligence, structured 3D latents, and contact-rich robotics.
- ai-for-scientific-discovery — AlphaEvolve, synthetic data for differentiable targets, and math/science discovery sources.
Link inventory
| Line | Link | Raw file | Fetch status |
|---|---|---|---|
| 21 | This New Tech Revolutionizes Biology… | Michael Levin | raw/articles/021-this-new-tech-revolutionizes-biology-michael-levin.md | fetched via oEmbed |
| 22 | Overview of Canada’s National Artificial Intelligence Strategy: AI for All | raw/articles/022-overview-of-canada-s-national-artificial-intelligence-strategy-ai-for-all.md | fetched |
| 23 | AI with AMD ROCm on Ubuntu: your questions answered | Ubuntu | raw/articles/023-ai-with-amd-rocm-on-ubuntu-your-questions-answered-ubuntu.md | fetched |
| 24 | mem0 (@mem0ai) on X | raw/articles/024-mem0-mem0ai-on-x.md | fetched |
| 25 | Everything looks fine at 4-bit | raw/articles/025-everything-looks-fine-at-4-bit.md | fetched via oEmbed |
| 26 | Figure’s First Full HQ Tour: From the Lab to the Factory Floor | raw/articles/026-figure-s-first-full-hq-tour-from-the-lab-to-the-factory-floor.md | fetched via oEmbed |
| 27 | Moravec’s Paradox and the Robot Olympics | raw/articles/027-moravec-s-paradox-and-the-robot-olympics.md | fetched |
| 28 | Hugo | Substack | raw/articles/028-hugo-substack.md | fetched |
| 29 | stepfun-ai/Step-3.7-Flash-GGUF · Hugging Face | raw/articles/029-stepfun-ai-step-3-7-flash-gguf-hugging-face.md | fetched |
| 30 | Paper page - When Should Models Change Their Minds? Contextual Belief Management in Large Language Models | raw/articles/030-paper-page-when-should-models-change-their-minds-contextual-belief-managemen.md | fetched |
| 31 | How far behind are open models? — LessWrong | raw/articles/031-how-far-behind-are-open-models-lesswrong.md | fetched |
| 32 | GitHub - Luthiraa/TALOS-V2: hardware implementation of transformers running microgpt at 50k+ tkps | raw/articles/032-github-luthiraa-talos-v2-hardware-implementation-of-transformers-running-mic.md | fetched |
| 33 | DiffusionBlocks: Block-wise Neural Network Training via Diffusion Interpretation | raw/articles/033-diffusionblocks-block-wise-neural-network-training-via-diffusion-interpretat.md | fetched |
| 34 | GitHub - roboflow/rf-detr: RF-DETR is a real-time object detection and segmentation model architecture developed by Roboflow, SOTA on COCO, designed for fine-tuning. (ICLR 2026) | raw/articles/034-github-roboflow-rf-detr-rf-detr-is-a-real-time-object-detection-and-segmenta.md | fetched |
| 35 | Local Coding Agents on Strix Halo and R9700: Pi, Opencode, and SWE-bench Mini Benchmarks | raw/articles/035-local-coding-agents-on-strix-halo-and-r9700-pi-opencode-and-swe-bench-mini-b.md | fetched via oEmbed |
| 36 | Components of A Coding Agent | raw/articles/036-components-of-a-coding-agent.md | fetched |
| 37 | But what exactly are world models? | raw/articles/037-but-what-exactly-are-world-models.md | fetched via oEmbed |
| 38 | Devansh in Artificial Intelligence Made Simple | raw/articles/038-devansh-in-artificial-intelligence-made-simple.md | fetched |
| 39 | arxiv.org/pdf/2605.19407 | raw/articles/039-arxiv-org-pdf-2605-19407.md | fetched |
| 40 | cdn.openai.com/pdf/74c24085-19b0-4534-9c90-465b8e29ad73/unit-distance-remarks.pdf | raw/articles/040-cdn-openai-com-pdf-74c24085-19b0-4534-9c90-465b8e29ad73-unit-distance-remark.md | fetched |
| 41 | AMD Powers Next-Generation Agent Computers with New Ryzen AI Halo Developer Platform and Ryzen AI Max PRO 400 Series Processors | raw/articles/041-amd-powers-next-generation-agent-computers-with-new-ryzen-ai-halo-developer.md | fetched |
| 42 | Lisan al Gaib (@scaling01) on X | raw/articles/042-lisan-al-gaib-scaling01-on-x.md | fetched |
| 43 | Using LLM in the shebang line of a script | raw/articles/043-using-llm-in-the-shebang-line-of-a-script.md | fetched |
| 44 | Principles and Practice of Deep Representation Learning | raw/articles/044-principles-and-practice-of-deep-representation-learning.md | fetched |
| 45 | We Ran 250 AI Agent Evals to Find Out if Skills Beat Docs. The Answer Is More Complicated Than We Expected | raw/articles/045-we-ran-250-ai-agent-evals-to-find-out-if-skills-beat-docs-the-answer-is-more.md | fetched |
| 46 | Quantized Neural Networks: The Only Guide You Need | raw/articles/046-quantized-neural-networks-the-only-guide-you-need.md | fetched |
| 47 | A Visual Guide to Gemma 4 | raw/articles/047-a-visual-guide-to-gemma-4.md | fetched |
| 48 | Accelerating Gemma 4: faster inference with multi-token prediction drafters | raw/articles/048-accelerating-gemma-4-faster-inference-with-multi-token-prediction-drafters.md | fetched |
| 49 | AlphaEvolve: A coding agent for scientific and algorithmic discovery | raw/articles/049-alphaevolve-a-coding-agent-for-scientific-and-algorithmic-discovery.md | fetched |
| 50 | Reddit - Please wait for verification | raw/articles/050-reddit-please-wait-for-verification.md | fetched |
| 51 | Reddit - Please wait for verification | raw/articles/051-reddit-please-wait-for-verification.md | fetched |
| 52 | GitHub - microsoft/TRELLIS.2: Native and Compact Structured Latents for 3D Generation | raw/articles/052-github-microsoft-trellis-2-native-and-compact-structured-latents-for-3d-gene.md | fetched |
| 53 | Efficient Video Intelligence in 2026 | raw/articles/053-efficient-video-intelligence-in-2026.md | fetched |
| 54 | In the AI Era, Shopify Is Investing in Junior Engineers—Not Cutting Them - CoderPad | raw/articles/054-in-the-ai-era-shopify-is-investing-in-junior-engineers-not-cutting-them-code.md | fetched |
| 55 | Neural Garbage Collection: Learning to Forget while Learning to Reason | raw/articles/055-neural-garbage-collection-learning-to-forget-while-learning-to-reason.md | fetched |
| 56 | Vision Banana | Google DeepMind | raw/articles/056-vision-banana-google-deepmind.md | fetched |
| 57 | Ask The Experts #2 – Lemonade: Fast GenAI on Ryzen AI and Radeon | raw/articles/057-ask-the-experts-2-lemonade-fast-genai-on-ryzen-ai-and-radeon.md | fetched via oEmbed |
| 58 | https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f | already ingested: 01-llm-wiki-github.md | skipped duplicate already in wiki |
| 59 | lmstudio-community/Qwen3-Coder-Next-GGUF · Hugging Face | raw/articles/059-lmstudio-community-qwen3-coder-next-gguf-hugging-face.md | fetched |
| 60 | How we OCR’ed 30,000 papers using Codex, open OCR models and Jobs | raw/articles/060-how-we-ocr-ed-30-000-papers-using-codex-open-ocr-models-and-jobs.md | fetched |
| 61 | Neural Computers | raw/articles/061-neural-computers.md | fetched |
| 62 | Synthetic Data for any Differentiable Target | raw/articles/062-synthetic-data-for-any-differentiable-target.md | fetched |
| 63 | Maziyar PANAHI (@MaziyarPanahi) on X | raw/articles/063-maziyar-panahi-maziyarpanahi-on-x.md | fetched |
| 64 | Manim Community | raw/articles/064-manim-community.md | fetched |
| 65 | Consulting the Oracle: Claude on the Future of Data | raw/articles/065-consulting-the-oracle-claude-on-the-future-of-data.md | fetched |
| 66 | How to Run Local LLMs with Claude Code | Unsloth Documentation | raw/articles/066-how-to-run-local-llms-with-claude-code-unsloth-documentation.md | fetched |
| 67 | Day 0 Support for Gemma 4 on AMD Processors and GPUs | raw/articles/067-day-0-support-for-gemma-4-on-amd-processors-and-gpus.md | fetched |
| 68 | Gemma 4: Byte for byte, the most capable open models | raw/articles/068-gemma-4-byte-for-byte-the-most-capable-open-models.md | fetched |
| 69 | Bonsai - a prism-ml Collection | raw/articles/069-bonsai-a-prism-ml-collection.md | fetched |
| 70 | Release 4.0.0 · huggingface/transformers.js | raw/articles/070-release-4-0-0-huggingface-transformers-js.md | fetched |
| 71 | GitHub - google-research/timesfm: TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting. | raw/articles/071-github-google-research-timesfm-timesfm-time-series-foundation-model-is-a-pre.md | fetched |
| 72 | Reddit - Please wait for verification | raw/articles/072-reddit-please-wait-for-verification.md | fetched |
| 73 | DexDrummer: In-Hand, Contact-Rich, and Long-Horizon Dexterous Robot Drumming | raw/articles/073-dexdrummer-in-hand-contact-rich-and-long-horizon-dexterous-robot-drumming.md | fetched |
| 74 | www.nature.com/articles/s41586-026-10265-5.pdf | raw/articles/074-www-nature-com-articles-s41586-026-10265-5-pdf.md | fetched |
| 75 | TurboQuant: Redefining AI efficiency with extreme compression | raw/articles/075-turboquant-redefining-ai-efficiency-with-extreme-compression.md | fetched |
| 76 | LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels | raw/articles/076-leworldmodel-stable-end-to-end-joint-embedding-predictive-architecture-from.md | fetched |
| 77 | arxiv.org/pdf/2603.19312v1 | raw/articles/077-arxiv-org-pdf-2603-19312v1.md | fetched |
| 78 | Claude Cowork: Claude Code power for knowledge work | Claude by Anthropic | raw/articles/078-claude-cowork-claude-code-power-for-knowledge-work-claude-by-anthropic.md | fetched |
| 79 | AI is redrawing the database market | ClickHouse | raw/articles/079-ai-is-redrawing-the-database-market-clickhouse.md | fetched |
| 80 | A Visual Guide to Attention Variants in Modern LLMs | raw/articles/080-a-visual-guide-to-attention-variants-in-modern-llms.md | fetched |
| 81 | Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI | raw/articles/081-skill-issue-andrej-karpathy-on-code-agents-autoresearch-and-the-loopy-era-of.md | fetched via oEmbed |
| 82 | Consciousness in humans and other things with Anil K Seth | The Royal Society | raw/articles/082-consciousness-in-humans-and-other-things-with-anil-k-seth-the-royal-society.md | fetched via oEmbed |
| 83 | Introducing Unsloth Studio | Unsloth Documentation | raw/articles/083-introducing-unsloth-studio-unsloth-documentation.md | fetched |
| 84 | SQLite FTS5 Extension | raw/articles/084-sqlite-fts5-extension.md | fetched |
| 85 | Run OpenClaw Locally On AMD Ryzen AI Max+ And Radeon GPUs | raw/articles/085-run-openclaw-locally-on-amd-ryzen-ai-max-and-radeon-gpus.md | fetched |
| 86 | GEPA: optimize_anything: A Universal API for Optimizing any Text Parameter | raw/articles/086-gepa-optimize-anything-a-universal-api-for-optimizing-any-text-parameter.md | fetched |
| 87 | Hermes Agent Documentation | Hermes Agent | raw/articles/087-hermes-agent-documentation-hermes-agent.md | fetched |
| 88 | Anthropic Courses | raw/articles/088-anthropic-courses.md | fetched |
| 89 | Bryan Catanzaro (@ctnzr) on X | raw/articles/089-bryan-catanzaro-ctnzr-on-x.md | fetched |
| 90 | Strix Halo GPU Benchmarks | raw/articles/090-strix-halo-gpu-benchmarks.md | fetched |
| 91 | GitHub - microsoft/markitdown: Python tool for converting files and office documents to Markdown. | raw/articles/091-github-microsoft-markitdown-python-tool-for-converting-files-and-office-docu.md | fetched |
| 92 | Applying Statistics to LLM Evaluations | raw/articles/092-applying-statistics-to-llm-evaluations.md | fetched |
| 93 | How is hardware reshaping LLM design? | raw/articles/093-how-is-hardware-reshaping-llm-design.md | fetched via oEmbed |
| 94 | THE 2028 GLOBAL INTELLIGENCE CRISIS | raw/articles/094-the-2028-global-intelligence-crisis.md | fetched |
| 95 | Reasoning Models Generate Societies of Thought | raw/articles/095-reasoning-models-generate-societies-of-thought.md | fetched |
| 96 | GitHub - open-webui/open-terminal: A computer you can curl ⚡ | raw/articles/096-github-open-webui-open-terminal-a-computer-you-can-curl.md | fetched |
| 97 | vLLM Triton Attention Backend Deep Dive | raw/articles/097-vllm-triton-attention-backend-deep-dive.md | fetched |
| 98 | AI Enabled Software Development and Jevons Paradox - Mike Grouchy | raw/articles/098-ai-enabled-software-development-and-jevons-paradox-mike-grouchy.md | fetched |
| 99 | Complete DSPy Tutorial - Master LLM Prompt Programming in 8 amazing examples! | raw/articles/099-complete-dspy-tutorial-master-llm-prompt-programming-in-8-amazing-examples.md | fetched via oEmbed |
| 100 | llama.cpp with ROCm | raw/articles/100-llama-cpp-with-rocm.md | fetched |
Notes
Social, Reddit, PDF, and some JavaScript-heavy pages may have only partial public metadata/text captures. The raw files preserve URL, final URL, line number, fetch status, and a text excerpt when available.