Channel Avatar

Vinh Nguyen

1.5K subscribers

Hi, I'm Vinh Nguyen (@vinhnx), a learn-by-doing software engineer exploring AI, machine learning, ...


09:53
Reverse Engineering and Rewriting E-Scooter Firmware in Rust
05:37
The Age of Agency: How Gemini 4 Neon Actually Works
38:57
Podcast \ How Gemini 4 Neon Rewrites AI Economics
07:56
What Makes Gemini 4 Neon Production-Ready
06:29
Video Special \ The Escalation Framework: GPT-6.2 Astra & Sol
01:04:05
Podcast \ How Software Sandboxes Restrict Access
01:05:46
Podcast \ GPT-6.2 Astra & Sol: Capabilities and Security Risks
08:17
GPT-6.2 Astra & Sol: How OpenAI's Control Room Stops Rogue AI
09:44
Software Sandboxing: The Basics
01:00:53
Podcast \ The AI Rich Get Richer Trap
12:10
The AI Rich Get Richer Trap
46:20
Podcast \ Nora: Normalized Orthogonal Row Alignment for Scalable Matrix Optimization
10:08
The Nora Optimizer
41:10
Podcast: How Gemini 4 Hydrogen Powers Agentic AI
06:37
Introducing Gemini 4 Hydrogen
08:34
Gemini 4 Hydrogen
40:57
Podcast \ LLMs Develop Novel Biases
06:01
Anatomy of Electronic Mail
09:09
LLMs Develop Novel Biases
05:34
The 2026 AI Illusion: Why Leaked Benchmarks Don't Matter for You
23:22
Why 2026 AI Evals Are Impossible
09:24
2026 AI Landscape Forecast
52:00
Podcast \ How Smart BMS Predicts True EV Battery Life
41:20
Podcast \ The $13,500 Summer Camp That Invented AI
11:26
No Silver Bullet
06:00
Inside Gemini 4 Carbon: The Heavy Machinery of AI
52:23
Podcast: How Gemini 4 Carbon Chains Cyber Exploits
06:41
Gemini 4 Carbon Replaces Chatbots With Swarms
54:44
Podcast: Claude Fable 5.2 built an operating system from scratch
05:59
Claude Fable 5.2: The Autonomy Threshold
09:41
Claude Fable & Claude Mythos 5.2: Power & Control
09:29
Fixing Range Anxiety: The Hybrid Physics ML Battery Architecture
08:18
The 1955 Blueprint: Deriving the Architecture of AGI
01:00:09
Podcast: How 1980s Tech Still Powers Email
08:32
The Hidden Mechanics of Email
10:41
How Smart BMS Predicts True EV Battery Life
50:41
Podcast \ HySparse2: Efficient Long-Context Retrieval via Two-Level KV Sharing
11:22
HySparse2: Efficient Long-Context Retrieval via Two-Level KV Sharing
07:02
Video Special \ The Efficiency Paradox: Dissecting Xiaomi's HySparse2
09:30
mold: Parallel Linker
01:07:05
Podcast \ mold: A Massively Parallel Linker Architecture
44:26
Podcast \ Harness Tax: The Impact of Software Scaffolding on Coding Costs
10:18
Harness Tax: The Impact of Software Scaffolding on Coding Costs
06:29
The Modeler's Lens: Translating Reality into Graph Theory
47:58
Podcast \ Terminal-Universe: Reconstructing Executable Environments from Agent Trajectories
49:27
Podcast: How AI Natively Generates Media
05:55
The Omnimodal Engine: A Visual Guide to Multimodal Generation
10:24
A Visual Guide to Multimodal Generation
11:06
Terminal-Universe: Reconstructing Executable Environments from Agent Trajectories
39:31
Podcast | Kimi Linear: AI Scaling
55:44
How Graph Theory Solves Complex Puzzles
07:39
Start Doing Graph Theory
22:11
Podcast | How Attention Residuals Fix LLM Memory
45:13
Google Gemini 4 Carbon Safety Report
06:20
Anatomy of a Frontier Risk: Inside Gemini 4 Carbon
09:33
Gemini 4: Shield & Sword
58:38
Podcast | How Pure Math Powers Modern Data Science
08:25
The Physics of AI-Native Work
06:41
Foundations of Large Language Model Training and Architecture
43:43
Podcast | How MAT Fills Massive Image Holes