Tech2Wild
@Tech2Wild

🎮 Tech, gaming, AI, and everything in between. 🤖 Building with it, not just talking about it. 🔥 From the mind of @ToNYD2WiLD

github.com/tonyd2wild
Joined March 2026
665 Photos and videos
🥊 NVFP4 vs EXL3, GLM-5.3-Flash on 2x DGX Spark. Same minute, same prompts, every number public. Here's what actually held up. 📊⚡️
Article

NVFP4 vs EXL3 on DGX Spark: same model, same boxes, same minute

GLM-5.3-Flash, two 4-bit quants, two 2-node lanes of DGX Spark, benched simultaneously. Decode c1 to c6, time to first token on fresh prompts, prefill from 1.6K to 211K tokens, prefix caching, mixed

8
3
41
7,730
Article

Two MiniMax H3 instances on two DGX Sparks, with a 1M-context Deepseek V4 Flash

Two DGX Sparks, one 1M-context LLM, and two video models. All at once. I run DeepSeek-V4-Flash across two DGX Sparks. Full 1,048,576-token context, a 1,473,052-token KV cache, serving my agents.

8
4
47
8,726
Article

How I Got Claude Opus 4.6 Back Into My OpenClaw Stack On My Existing Max Plan

How I Got Claude Opus 4.6 Back Into My OpenClaw Stack (For Free, On My Existing Max Plan) If you're running OpenClaw and got hit by Anthropic's recent ban on using Claude Max plan OAuth, this article
1
144
Article

My AI agent never forgets a conversation again. Here's how we built it tonight

THE PROBLEM I run 8 AI agents. Kai handles strategy. Donnie runs operations. Cruz does engineering. Maya handles intel. They talk to me on Telegram, Discord, and through a voice system called Eyes of

1
97
Load more