https://ai.ericeaglstun.com/Recent content in Home on Eric Eaglstun · AIHugoen-ushttps://ai.ericeaglstun.com/glossary/agi/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/agi/Hypothetical AI that matches humans across essentially all intellectual tasks — a contested, moving goalpost.https://ai.ericeaglstun.com/glossary/attention/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/attention/How tokens weigh each other (query · key · value) — the heart of the transformer.https://ai.ericeaglstun.com/glossary/cuda/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/cuda/NVIDIA’s GPU-compute platform; the default ML backend.https://ai.ericeaglstun.com/glossary/cudnn-cublas/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/cudnn-cublas/NVIDIA’s CUDA math & deep-learning libraries.https://ai.ericeaglstun.com/glossary/dimensions/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/dimensions/Independent axes of variation; ML vectors live in hundreds or thousands of them.https://ai.ericeaglstun.com/glossary/embeddings/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/embeddings/Dense vectors where distance & direction encode meaning — the backbone of semantic search and RAG.https://ai.ericeaglstun.com/glossary/gan/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/gan/Generator-vs-discriminator generative architecture.https://ai.ericeaglstun.com/glossary/ggml/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/ggml/C/C++ tensor library powering llama.cpp; runs GGUF.https://ai.ericeaglstun.com/glossary/gguf/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/gguf/llama.cpp’s single-file format for quantized local LLMs.https://ai.ericeaglstun.com/glossary/gpt/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/gpt/Generative pre-trained (decoder-only) transformer LLM.https://ai.ericeaglstun.com/glossary/latent-space/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/latent-space/The hidden vector space where geometry encodes meaning.https://ai.ericeaglstun.com/glossary/llamacpp-vs-ollama/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/llamacpp-vs-ollama/The local-inference engine vs the wrapper built on it.https://ai.ericeaglstun.com/glossary/lora/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/lora/Low-rank adapters; parameter-efficient fine-tuning.https://ai.ericeaglstun.com/glossary/machine-learning/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/machine-learning/Systems that learn patterns from data instead of hand-written rules.https://ai.ericeaglstun.com/glossary/metal/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/metal/Apple’s GPU-compute API; powers MLX and Metal-backed ML.https://ai.ericeaglstun.com/glossary/mlx/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/mlx/Apple-silicon ML framework; the Mac answer to GGUF.https://ai.ericeaglstun.com/glossary/mps/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/mps/Metal Performance Shaders; Apple’s cuDNN-equivalent ops.https://ai.ericeaglstun.com/glossary/parameters/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/parameters/Learned weights; the count is model size & memory cost.https://ai.ericeaglstun.com/glossary/tensor/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/tensor/N-dimensional numeric array; the core ML data structure.https://ai.ericeaglstun.com/glossary/transformer/Mon, 01 Jan 0001 00:00:00 +0000https://ai.ericeaglstun.com/glossary/transformer/The self-attention architecture behind modern LLMs.