bycloud
bycloud · 269 個帶字幕的影片bycloud 在 WinkLingo 收錄了 269 個影片(這個頻道有 24萬 個訂閱),每個影片都能逐字檢索。
查看 YouTube 原頻道英文簡介
Frontier AI Research Explained
收錄的影片
19:19DeepSeek 新 LLM 引擎讓 GPU 價值提升 3 倍
DeepSeek's New LLM Engine Just Made GPUs 3x More Valuable
20:12目前最具有創意的 AI 沙箱逃逸
The Most Creative AI Sandbox Escape Yet
19:30Muon:LLM 領域最被低估的突破
Muon: The Most Underrated Breakthrough in LLMs
14:56小米向我們展示了百萬美元級別的 RL 訓練是什麼樣子
Xiaomi Just Showed Us What Million-Dollar RL Training Looks Like
25:04DeepSeek-V4.1 Flash:迄今為止最瘋狂的最佳化
DeepSeek-V4.1 Flash: The Most Insane Optimization So Far
17:51決策模型:不生成文字的模型如何變得更強大?
Decision Models: How Can A Model Be Better Without Generating Words?
23:02LLM 推理即服務 (Inference-as-a-Service) 的奇特經濟學
The Strange Economics of LLM Inference-as-a-Service
19:52NVIDIA 讓 4-Bit AI 運行了起來,但這簡直是一場噩夢
NVIDIA Made 4-Bit AI Work. It’s a Nightmare.
24:30Neocloud 是下一個網際網路泡沫嗎?詳解 Neocloud 經濟學
Are Neoclouds the Next Dot-Com Bubble? Neocloud Economics Explained
15:41機器人迎來了它們的 GPT-3 時刻
Robots Just Had Their GPT-3 Moment
15:33他們找到了竊取前沿大模型(LLM)推理能力的方法
They Found a Way to Steal Frontier LLM’s Reasoning
16:08訓練大語言模型最荒誕的方式……記憶體竟然減少了3倍!?
The Most Absurd Way To Train LLMs... With 3x Less Memory!?
13:59LongCat 2.0:NVIDIA 護城河終結的開始?
LongCat 2.0: The Beginning of the End of NVIDIA MOAT?
17:27深度解析 Kimi K3 驚人的架構突破
A Closer Look At Kimi K3's INSANE Architecture Breakthrough
15:57GLM-5.2:DeepSeek 對 RL 的看法錯了?
GLM-5.2: DeepSeek Was Wrong About RL?
16:23DSpark:詳解 DeepSeek-V4 驚人的計算最佳化
DSpark: DeepSeek-V4's Insane Compute Optimization Explained
2:55GPT-5.6 Sol 與基準最小化?? #LLM #ai #chatgpt
GPT-5.6 Sol & Benchmark Minimizing?? #LLM #ai #chatgpt
17:43微軟剛剛揭秘了 LLM 前沿資料工程的秘密
Microsoft Just Dropped LLM's Frontier Data Engineering Secrets
18:38DeepSeek 被刪除的視覺論文簡直太瘋狂了...
DeepSeek's Deleted Vision Paper Is Nuts...
19:10迴圈 LLM?詳解遞迴深度 Transformer (Recurrent Depth Transformer)
Looped LLM? Recurrent Depth Transformer Explained
22:01一家中國手機公司是如何在開源大模型中登頂的?
How Did A Chinese Phone Company Topped Open Source LLM?
26:44DeepSeek V4 瘋狂的基礎設施設計
The Insane Infrastructure Design of DeepSeek V4
21:13DeepSeek 是如何把 V4 做得這麼便宜的?
How Did DeepSeek Make V4 So Cheap?
13:58Attention Residuals:Kimi AI 優雅的 LLM 架構突破
Attention Residuals: Kimi AI's Elegant LLM Architecture Breakthrough