My ASUS AI Agent Computer: Local LLMs. Coding & Agent Demos

https://www.asus.com/microsite/agent-computer/in/?utm_source=official-site&utm_medium=asus-hero-banner&utm_campaign=Powering-Every-AI-Choice Can you run LLMs, write code, and build AI agents on your own workstation? In this video, I walk you through my ASUS AI Agent Computer and demonstrate local AI workflows—from running models with Ollama to comparing a local research agent against a cloud-based model. We explore GPU utilization, VRAM, response times, tool calls, and estimated API costs through practical examples. I also demonstrate ASUS Zenni Claw for everyday productivity, including an AI news digest and a slide-generation workflow. If you’re learning Agentic AI, building AI applications, or exploring a workstation for local inference, this walkthrough connects the hardware with real development tasks. **⏱️ TIMESTAMPS & TOPICS** 00:00 Introducing my AI workstation 00:22 Why run AI locally? Local, cloud, and hybrid approaches 01:57 ASUS workstation hardware tour 03:25 Ollama installation and local model setup 04:06 Running and chatting with Qwen 3.5 locally 06:31 Finding and running additional models with Ollama 07:09 Generating Python code for a Tetris game with Llama 3.2 07:34 Checking GPU utilization, VRAM, and NVIDIA-SMI 08:09 Research agent project walkthrough with Claude Code 09:26 Streamlit app setup and model selection 10:07 Local vs cloud agents: facts and calculation task 11:18 Comparing response times, tool calls, and estimated costs 12:11 Live AI news research: local vs cloud 13:17 Timeouts, execution limits, and cost observations 14:00 Compare-and-compute task and local model considerations 15:44 Introducing ASUS Zenni Claw 16:19 Local models, additional providers, and API configuration 16:46 Work Assistant demo: generating an AI news digest 18:28 Exploring productivity skills and workflows 18:51 Trying slide generation about Agentic AI systems 20:17 Key takeaways: AI agents and the infrastructure behind them 20:54 Choosing between local, cloud, and hybrid AI 21:07 Workstation resources and closing thoughts 21:20 Would you run your AI agents locally? **💡 WHAT YOU’LL LEARN** - How to get started with local LLMs using Ollama. - How GPU memory and model size affect local inference. - How to generate code using a locally running model. - How to monitor GPU utilization with Task Manager and NVIDIA-SMI. - How a Streamlit research agent compares local and cloud models. - How to inspect response time, tool usage, estimated API costs, and failures. - How ASUS Zenni Claw approaches everyday agentic workflows. - Where local, cloud, and hybrid AI can fit into your projects. The comparisons shown are observations from these specific models, prompts, and settings. Local inference avoids per-token cloud inference charges for the locally executed work, while hardware, electricity, and connected services still have costs. **🖥️ HARDWARE FEATURED** - Motherboard: [ROG Maximus Z890 Hero](https://rog.asus.com/in/motherboards/rog-maximus/rog-maximus-z890-hero/) - GPU: [ROG Strix GeForce RTX 5070 Ti — 16GB GDDR7](https://rog.asus.com/in/graphics-cards/graphics-cards/rog-strix/rog-strix-rtx5070ti-16g-gaming/) - Power supply: [ASUS TUF Gaming 1000P](https://www.asus.com/in/motherboards-components/power-supply-units/tuf-gaming/tuf-gaming-1000p/) - AIO cooler: [ROG Strix LC III 360 ARGB](https://rog.asus.com/in/cooling/cpu-liquid-coolers/rog-strix-lc/rog-strix-lc-iii-360-argb/) **🔗 USEFUL LINKS** - [ASUS Agent Computer](https://www.asus.com/microsite/agent-computer/in/) - [ASUS Agent Computer at EliteHubs](https://elitehubs.com/pages/asus-agent-computer) - [ASUS Zenni Claw — features, download, and device compatibility](https://www.asus.com/content/asus-zenni-claw/) - [Ollama](https://ollama.com/) **💬 JOIN THE DISCUSSION** Would you run your AI agents locally, use cloud APIs, or combine both? Share your preferred setup and the workflows you would build in the comments. Like the video and subscribe for more practical tutorials on Agentic AI, LLMs, and AI application development. #AgenticAI #LocalLLM #ASUS #Ollama #AIAgents