Based on this YouTube video's total view count of 18.3K views and industry-standard rates, the estimated total earning is $13 - $37 through ad revenue. Historical data is not yet available to calculate daily, weekly, or monthly averages.

Explore Qwen 3.8 27B Teardown: They Deleted 75% of the Attention with 18,321 views, 357 likes, and 42 comments. Experience the impact of this video content that has captured audience attention.

Why did Alibaba build a 27-billion parameter model where 48 of its 64 layers run Gated DeltaNet linear attention instead of traditional transformer attention? Meet Qwen 3.8 27B and the 3:1 architectural ratio replacing transformers. 🔔 Subscribe: https://www.youtube.com/channel/UC0DZj1PNa_Fp0MU6uPSKv5w?sub_confirmation=1 💙 Become a Member: https://www.youtube.com/channel/UC0DZj1PNa_Fp0MU6uPSKv5w/join 🐦 Twitter/X: https://x.com/cloud_codes 💬 Discord: https://discord.gg/4kJqEBMMf In this deep dive, Cloud Codes breaks down the computer architecture, linear algebra, and KV cache memory arithmetic behind Qwen 3.8 27B Dense. We examine its 64-layer hybrid layout (48 Gated DeltaNet layers and 16 Gated Attention layers), analyze how updating a running state matrix instead of storing an expanding KV cache shoebox cuts 1M token context memory from 244 GB down to 61 GB, and review its Apache 2.0 open weights. Furthermore, we audit company-reported benchmarks showing Qwen 3.8 27B beating Claude Opus 4.6 Max on SWE-bench Pro (61.7 vs 53.4) and OSWorld (84.3 vs 72.7), expose the 9-point loss on Humanity's Last Exam (HLE), dissect the model card's Claude Code harness footnote, analyze Simon Willison's 17GB GGUF local run on an M5 Max MacBook Pro, and evaluate why linear attention hybrids represent the future of local LLM serving. If this helped you understand backend architecture, system design, and how to build faster software, subscribe to Cloud Codes for a new infrastructure breakdown every single week! Build, solve, deploy. 🔗 Repositories & Sources Mentioned: Qwen 3.8 27B Official Model: https://huggingface.co/Qwen/Qwen3.8-27B Unsloth Qwen 3.8 27B GGUF Quants: https://huggingface.co/unsloth/Qwen3.8-27B-GGUF Gated Delta Networks Architecture Paper: https://arxiv.org/abs/2412.06464 SGLang Inference Engine: https://github.com/sgl-project/sglang Hacker News — Qwen 3.8 Discussion: https://news.ycombinator.com/item?id=48966120 ⏱️ Video Chapters: 0:00 - The 75% Non-Transformer Architecture 1:01 - The Memory Wall: Why Attention KV Caches Explode 1:43 - The 61GB vs 244GB KV Cache Arithmetic 3:07 - What Is Gated DeltaNet? (Shoebox vs Running Balance) 4:44 - The 3:1 Ratio: Interleaving Linear Attention & Full Attention 5:20 - Benchmarks: Qwen 27B vs Claude Opus 4.6 Max 6:41 - Where Qwen Loses: Terminal-Bench & HLE Reasoning 8:31 - Dense 27B Footprint & MTP Speculative Decoding 9:47 - The Footnote: Who Ran These Benchmarks? 11:19 - Local Developer Testing: Simon Willison's 17GB GGUF Run 12:45 - The Architecture Shift: Qwen 3.8 & Nemotron 3.5 13:47 - Final Verdict: Why the Architecture Won #qwen #qwen38 #gateddeltanet #localai #systemdesign #cloudcodes #machinelearning #transformers #python #softwareengineering User Queries: qwen 3 8 27b architecture gated deltanet linear attention why 75 percent of qwen 3 8 27b is not a transformer qwen 3 8 27b 1m context kv cache 61gb vs 244gb gated deltanet vs mamba 2 state space models simon willison qwen 3 8 27b lm studio m5 max benchmark qwen 3 8 27b swebench pro osworld vs claude opus yarn factor 4 0 context extension qwen 3 8 multi token prediction mtp head qwen 3 8 27b qwen 3 8 27b apache 2 0 open weights cloud codes qwen 3 8 27b efficiency breakdown

With SocialCounts.org’s view counter, track your YouTube video’s live view count and YouTube likes count in real time with fast, reliable updates.

Watch every YouTube video live view count rise with our real-time YouTube views tracker—built for accuracy and minimal delay.

Follow YouTube real time views as they happen, using our dedicated view counter for YouTube videos.

Get up-to-date live view count on YouTube and see real-time growth with SocialCounts.org’s smart tracking tools.