There is a recurring “wall” every AI hobbyist hits when working with Novel View Synthesis (NVS). You generate a beautiful second view of a room, but as soon as you try to “walk” further into the scene, the geometry falls apart like a house of cards. Recently, I came across the paper “SceneCompleter: Dense 3D Scene … Read more
We’ve all seen the benchmarks. The new “Reasoning” models (like the o1 series or fine-tuned Llama-3 variants) claim to possess human-like logic. But after building my dual-RTX 4080 lab and running these models on bare-metal Ubuntu, I’ve started to see the cracks in the mirror. Is it true “System 2” thinking, or just an incredibly … Read more
Alright, fellow hardware junkies and algorithm enthusiasts. You know my journey: from building my dual-RTX 4080 rig to wrestling with scaling laws and even trying to birth a local data scientist with AutoMind. Each step has been about pushing the boundaries of what one person can do with local compute. But what if the next … Read more
After spending weeks obsessing over scaling laws and raw TFLOPS, I decided it was time to move up the stack. It’s one thing to have a powerful model; it’s another to have an Agent that knows how to use it. I took the architecture described in my recent overview of AutoMind AI Agent — an adaptive agent for automated … Read more