AI/ML Link Batch — 2026-06-27

This query page records the first 20 unlabeled links from links-no-label-aah-ai-gold-jjh-strix-2026-06-27.txt-first20.txt and the raw wiki files created for them.

Synthesized themes

• llm-wikis and agent-skills: durable agent knowledge and reusable skills.
• local-ai-benchmarks, diffusion-language-models, and small-reasoning-models: practical model performance on constrained or local hardware.
• reinforcement-learning-history: historical context around Q-learning and RL.
• vector-indexes: retrieval infrastructure.
• multimodal-model-tools: image/video/multimodal tooling and evaluation.

Link inventory

#	Link	Raw file	Fetch status
1	llm-wiki · GitHub	raw/articles/01-llm-wiki-github.md	fetched
2	Reddit - Please wait for verification	raw/articles/02-reddit-please-wait-for-verification.md	fetched
3	Reddit - Please wait for verification	raw/articles/03-reddit-please-wait-for-verification.md	fetched
4	pi-bench | Local AMD Benchmarks	raw/articles/04-pi-bench-local-amd-benchmarks.md	fetched
5	Robonaissance | Hugo | Substack	raw/articles/05-robonaissance-hugo-substack.md	fetched
6	archive.is/F0fyO	raw/articles/06-archive-is-f0fyo.md	fetch failed; HTTPError: HTTP Error 429: Too Many Requests
7	Automating SKILL.md Generation for Computer-Using Agents via Interaction Trajectory Mining	raw/articles/07-automating-skill-md-generation-for-computer-using-agents-via-interacti.md	fetched
8	Vik Paruchuri on X: “We’re open sourcing a 9B model that extracts structured data from documents at near-frontier performance. - 90.2% on our bench, vs Gemini 3.5 Flash at 91.3% - Leads extraction models like NuExtract3 (81.5%) - 9.5s p50 timings - Pass JSON schema https://t.co/zIdRbGjaY5” / X	raw/articles/08-vik-paruchuri-on-x-we-re-open-sourcing-a-9b-model-that-extracts-struct.md	fetched
9	VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models	raw/articles/09-vibethinker-3b-exploring-the-frontier-of-verifiable-reasoning-in-small.md	fetched
10	Midjourney on X: “A technical dive inside our new “Midjourney Scanner” https://t.co/wJBHz2O7ro” / X	raw/articles/10-midjourney-on-x-a-technical-dive-inside-our-new-midjourney-scanner-t-c.md	fetched
11	Running DiffusionGemma on AMD Strix Halo and Decade-Old Tesla P40s | TinyComputers.io	raw/articles/11-running-diffusiongemma-on-amd-strix-halo-and-decade-old-tesla-p40s-tin.md	fetched
12	Everything I Learned Training Frontier Small Models — Maxime Labonne, Liquid AI	raw/articles/12-everything-i-learned-training-frontier-small-models-maxime-labonne-liq.md	fetched via oEmbed
13	Yann LeCun’s $1B Bet Against LLMs [Part 2]	raw/articles/13-yann-lecun-s-1b-bet-against-llms-part-2.md	fetched via oEmbed
14	AINews: Loopcraft: The Art of Stacking Loops	raw/articles/14-ainews-loopcraft-the-art-of-stacking-loops.md	fetched
15	Lucas Beyer (bl16) on X: “You may have recently heard claims that video generation models are “dumb” about physics, and only “world models” (V-JEPA, specifically) have a valid internal model of physics. This turns out to be false. In a recent paper, researchers show that a LINEAR probe of diffusion https://t.co/uqleVq5lq2” / X	raw/articles/15-lucas-beyer-bl16-on-x-you-may-have-recently-heard-claims-that-video-ge.md	fetched
16	Daniel Han on X: “We also made a finetuning notebook showing how DiffusionGemma works well for solving Sudoku - normal LLMs think left to right, and can’t edit its previous responses, whilst diffusion LLMs can update its past tokens! This makes them super powerful for refinement tasks! https://t.co/O7OgbuvXRz” / X	raw/articles/16-daniel-han-on-x-we-also-made-a-finetuning-notebook-showing-how-diffusi.md	fetched
17	GitHub - RyanCodrai/turbovec: A vector index built on TurboQuant, written in Rust with Python bindings · GitHub	raw/articles/17-github-ryancodrai-turbovec-a-vector-index-built-on-turboquant-written.md	fetched
18	Hugo (@robonaissance): “Christopher Watkins was at Philips Research Labs in Surrey, working on expert systems. He had left Cambridge two years earlier without finishing his PhD. He had no paper to present at the workshop, and no specific agenda. During one of the talks at UC Irvine, he was bored. He i…”	raw/articles/18-hugo-robonaissance-christopher-watkins-was-at-philips-research-labs-in.md	fetched
19	Omar Sanseviero on X: “llama.cpp just added video input support 👀 You can now enjoy Gemma 4 video understanding capabilities in your chat completions endpoint and via mtmd-cli https://t.co/p79ZKMXLue” / X	raw/articles/19-omar-sanseviero-on-x-llama-cpp-just-added-video-input-support-you-can.md	fetched
20	Agent Skills Overview - Agent Skills	raw/articles/20-agent-skills-overview-agent-skills.md	fetched

Notes

Some social and archive pages were only partially captured because the public web versions are gated, rate-limited, or optimized for logged-in browsers. Their metadata was still preserved in raw/articles/ where available.