Attention — Oct 7, 2026
What caught my attention on Oct 7 (17 items).
“Cycular” Financing w/ GamersNexus
Better Offline (Podcast:04:36) - Published 2026-10-06 17:00 - Discovered 2026-10-07 04:36
CZM Rewind: The Man That Destroyed Google Search
Better Offline (Podcast:09:00) - Published 2024-11-26 17:00 - Discovered 2026-10-07 09:00
The Rot Society
Better Offline (Podcast:09:56) - Published 2024-11-19 17:00 - Discovered 2026-10-07 09:56
William Lazonick on How The Stock Market Killed Tech
Better Offline (Podcast:10:58) - Published 2024-11-12 17:00 - Discovered 2026-10-07 10:58
The Streisand Effect with Mike Masnick
Better Offline (Podcast:11:59) - Published 2024-11-05 17:00 - Discovered 2026-10-07 11:59
A Nobel Economist’s Plan To Tax Digital Ads
Better Offline (Podcast:12:53) - Published 2024-10-29 17:00 - Discovered 2026-10-07 12:53
A Requiem for Prabhakar Raghavan
Better Offline (Podcast:13:24) - Published 2024-10-24 17:00 - Discovered 2026-10-07 13:24
Embedding Gemma 2: Multimodal Search & RAG (Free Colab)
Prompt Engineering (Article:13:33) - Published 2026-10-06 18:00 - Discovered 2026-10-07 13:33
The video introduces Embedding Gemma 2, a compact open multimodal embedding model from Google DeepMind that maps text, code, images, video, and audio into a shared vector space. The presenter explains why this matters for search and retrieval-augmented generation: much real-world data is visual or auditory, and text-only RAG cannot retrieve photos, screen recordings, voice notes, or videos unless…
Running Qwen3.8-Flash-Next in 12GB of VRAM
NoPlaceLikeLocalhost (Article:13:35) - Published 2026-10-06 18:00 - Discovered 2026-10-07 13:35
In this video, the creator tests Strata, a tool that claims to run very large mixture-of-experts language models on systems with limited VRAM. The model examined is Qwen3.8-Flash-Next, a 125-billion-parameter MoE model with about 6 billion active parameters. In full precision it would require hundreds of gigabytes of memory, and even common quantizations remain far beyond typical consumer GPUs. Th…
“Cycular” Financing w/ GamersNexus | Better Offline
Better Offline (Article:13:56) - Published 2026-10-06 18:00 - Discovered 2026-10-07 13:56
In this episode of Better Offline, Ed Zitron and Steve Burke of GamersNexus discuss Nvidia’s “cyclical” financing and growing concerns that the AI boom is being sustained by circular deals, debt, and questionable data-center promises. They argue that Nvidia’s revenue is heavily concentrated among a few large buyers, many of whom may not have deployed the GPUs they purchased, raising fears of an AI…
Data Vampires with Paris Marx
Better Offline (Podcast:14:26) - Published 2024-10-22 17:00 - Discovered 2026-10-07 14:26
Zephyr Teachout On Corruption
Better Offline (Podcast:15:23) - Published 2024-10-15 17:00 - Discovered 2026-10-07 15:23
The Subprime AI Crisis
Better Offline (Podcast:16:29) - Published 2024-10-10 17:00 - Discovered 2026-10-07 16:29
Why OpenAI Is A Terrible Business
Better Offline (Podcast:17:35) - Published 2024-10-08 17:00 - Discovered 2026-10-07 17:35
Mythos: not the magical AI hacking tool after all
Pivot to AI (Article:18:23) - Published 2026-10-06 18:00 - Discovered 2026-10-07 18:23
In this Pivot to AI episode, David Gerard argues that Anthropic’s Claude Mythos, marketed in April as an exceptionally dangerous AI hacking tool, has not lived up to that reputation. Although Mythos was promoted as so effective that the US government temporarily restricted it, Gerard says its apparent success relied heavily on Anthropic’s subsidies and extensive human review. Anthropic’s Project G…
The Department of Justice V Google Ads: Part 4
Better Offline (Podcast:22:06) - Published 2024-10-01 17:00 - Discovered 2026-10-07 22:06
How To Break the Internet with Chris Stokel-Walker
Better Offline (Podcast:23:21) - Published 2024-09-26 17:00 - Discovered 2026-10-07 23:21