KV Caching in Attention - Unlocking Faster LLM Inference
Discover how KV caching dramatically speeds up text generation in Large Language Models by cleverly reusing previous calculations in the attention mechanism.
Read more →Blog Posts
Decoding Transformers
An in-depth look at how Transformer models work, covering pretraining, batching, embeddings, attention mechanisms, and sampling strategies for LLMs.
Read more →2 Startups
Startups aren't a single entity but a two-part journey: pre-PMF and post-PMF. This post explains why the distinction is critical and how advice often misaligns.
Read more →Distribution for PMF
Founders with distribution have an advantage in finding PMF. This post explores how, the common pitfalls, and a potential solution for leveraging audience effectively.
Read more →Students in Tech who stutter/stammer
An exclusive community for tech students who stutter/stammer. Learn why it's different, who's behind it, and how to join.
Read more →How LLM Tokenizers Work: Understanding Byte-Pair Encoding (BPE)
ChatGPT doesn't just 'understand' English; it processes numbers. This post explains how tokenizers, especially Byte-Pair Encoding (BPE), convert your text into a format LLMs can work with using a simplified example.
Read more →Okay, so it is a skill issue
My journey from initial disappointment with GPT-4 image generation to understanding the crucial role of skilled prompting. Includes an example with a detailed prompt.
Read more →What am I doing this summer.
Posting your goals to world keeps you super accountable
Read more →Sharing all the context I've gathered
If you're looking to learn how to build a startup, start here. This post covers essential startup and technical context for aspiring builders.
Read more →