Tag
#llm
Every story tagged llm, newest first.

Mixture-of-Experts Models: How They Work and Why They Cut Inference Costs
MoE models activate only a fraction of their parameters per token: DeepSeek-V3 fires 37B of 671B, delivering large-model quality at small-model compute cost. How routers, experts, and load balancing actually work, and where the savings are and are not real. With an interactive dense-vs-MoE view.
BitByteCore Silicon Desk · Aug 6, 2026 · 10 min read

How to Evaluate a Local LLM for a Real Task: A Repeatable Testing Framework
A leaderboard score is a population statistic, measured on a full-precision model you'll never run. Whether a local model can handle your invoice pipeline is a different question — here's the repeatable discipline for answering it on your own hardware.
BitByteCore AI Desk · Aug 5, 2026 · 11 min read

The best local LLM runners in 2026: Ollama, LM Studio, vLLM, and more
For most people the best local LLM runner in 2026 is still Ollama — free, cross-platform, out of your way. But LM Studio, vLLM, Apple MLX, llama.cpp, Jan, GPT4All, and Open WebUI each win a specific job. Here's which to pick — and what actually fits your GPU.
BitByteCore Research · Aug 4, 2026 · 9 min read

The best smartphones for on-device AI in 2026
For most people in mid-2026, the best on-device AI phone is the Samsung Galaxy S26 Ultra — it runs the widest range of AI tasks locally, no server round-trip, at a price you can justify. Here's how it compares to the iPhone 17 Pro Max, Pixel 10, OnePlus 13 and the ROG Phone 10 Pro.
Silicon Desk · Aug 3, 2026 · 11 min read

Model Distillation: How Small Models Learn to Punch Above Their Weight
Training a tiny model to mimic a giant one sounds like a compromise. It is actually a distinct training discipline, and done right it produces models that beat their parameter count in the ways that matter for shipping AI. With an interactive look at the dark knowledge a soft label carries.
BitByteCore Silicon Desk · Jul 28, 2026 · 9 min read
More stories
Article · aiRAG vs Fine-Tuning: Which One Your Use Case Actually NeedsJul 27, 2026 · 10 min read
Article · securityPrompt Injection: The Unsolved Security Hole in AI AppsJul 27, 2026 · 13 min read
Article · aiWhat an AI Agent Really Is: Stripping Away the HypeJul 27, 2026 · 14 min read
Article · aiHow a transformer model actually worksMay 13, 2026 · 4 min read
Article · aiThe real difference between training and inferenceMay 12, 2026 · 4 min read
Article · aiWhat a context window actually isMay 11, 2026 · 4 min read
Article · aiWhat RAG actually is and is notMay 10, 2026 · 4 min read