Poolside's 118B Laguna model beats rivals ten times its size on coding, trained in nine weeks with anti-reward-hacking guardrails. Also: clustering Ryzen AI boxes to run 400B models at home, and why top models still can't out-think hackers.

Source episodes:

- AOUSD Drives Global 3D Data Interoperability and Agentic AI Workflows with New Core Specification Milestones and Members — The Linux Foundation (Press Release) (Harold Web Research) https://www.linuxfoundation.org/press/aousd-drives-global-3d-data-interoperability-ai-workflows-with-core-spec-milestones-and-members?hs_amp=true

- How to Get the Most from AI This Summer (The AI Daily Brief) https://anchor.fm/s/f7cac464/podcast/play/123339068/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-6-26%2F428640174-44100-2-d9db74d5d310f.mp3

- Hacker News Show #10: mindwalk, spaceprojectsim, dom-docx, ctx, codealmanac, bramble, remux, leaves (Github Awesome) https://www.youtube.com/watch?v=DCCkBPcIVAo

- This AI Technology Will Replace Millions (Here's How to Prepare) (Nate Herk | AI Automation) https://www.youtube.com/watch?v=Ums8suyAG1A

- AMD Says 2 Ryzen AI Halos Can Run a 400B Model... I Tested It (Alex Ziskind) https://www.youtube.com/watch?v=FE1Uyhg5hjw

- Laguna S 2.1: The Best Local Agentic Coder? (Prompt Engineering) https://www.youtube.com/watch?v=H_Lbe69XO_8

- Training Frontier Models to Out-Think Hackers — Uri Rolls, Arithmetic & Thom Wolf, Hugging Face (AI Engineer) https://www.youtube.com/watch?v=O-CBZ3JtRvo

- Inkling: This Open-Weight Model Wants To Be Fine Tuned (Better Stack) https://www.youtube.com/watch?v=YAcIUCjsRnU

- Jack Dorsey's New App Is Coming for Slack and GitHub (buzz) (Better Stack) https://www.youtube.com/watch?v=CHEMPZ87FLw