Part 2 of 4
Inside System 1 Models
How a transformer stops writing text and starts providing calibrated decisions.
Harkanwal Bedi
Staff Forward Deployed AI Engineer @ Google Seattle, WA
Posts
All on Medium-
-
Part 1 of 4
Thinking, Fast and Slow for Production Agents
Pairing open-weights System 1 decision models with an LLM so every model plays to its strengths.
Open source
All projects-
Merged C++
ggml-org/
llama.cpp #30779 llama-server now visits parent slots before their child slots, so requests that share a prompt prefix no longer need a third forward pass. Released in b12188.
-
Merged Python
AI-Hypercomputer/
maxdiffusion #476 Skips the CUDA-only nvidia-cuda-cccl dependency on macOS, where no wheels exist, so local installs no longer fail on Apple Silicon.