Latest / AGI Dreams – Open, Uncensored, & Local - AI News Digest
AGI Dreams Podcast – February 20, 2026
Hardware's Shifting Fault Lines
In this episode:
Hardware's Shifting Fault Lines — NVIDIA's fifteen-year GPU market segmentation via FP64 precision is collapsing as AI training moves to FP16/FP8/FP4. The Blackwell B300 sacrifices FP64 for tensor cores, while a scaled-back $30B NVIDIA-OpenAI deal signals capital recalibration across the industry.
Local AI Comes of Age — The llama.cpp team joins Hugging Face to build seamless local inference integration, while an enterprise team runs an 80B-parameter agentic coding assistant on $8,000 of DGX Spark hardware with zero API costs and…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.