LLMs, from the Beginning — Part 1: The Next Token
See how next-token completion becomes few-shot prompting, chat, tool use, and reasoning through interactive, token-level exhibits.
Posts on Coding in Space
Recent content in Posts on Coding in Space
Latest posts
LLMs, from the Beginning — Part 2: Remembering the Prompt
Open the KV cache, count what every remembered token costs, and explore how modern attention designs handle long context.
LLMs, from the Beginning — Part 3: Breaking the One-Token Barrier
Connect arithmetic intensity, continuous batching, sparse experts, and speculative decoding to the latency each user feels.
Cipher – A Vision for the Future of Tab Completion
Tab-completion LLMs (large language models) are models that provide real-time suggestions while a developer types code, often integrated into editors ranging from Vim to IntelliJ as plugins, but sometimes as an integral part of the editor.
Phi-3 is convinced that Microsoft made ChatGPT
I recently noticed a strange quirk of Phi-3.
Zapper: A Very Fast Templating Engine
Thirty times faster than Handlebars, half the features!
Garbage Collectors
I recently gave a five minute “lightning talk” on what garbage collection is and how it works in modern languages.