Hi, I'm Newvick. I write about tech and things I find interesting.
Skip lists are used in redis and LSM-trees, delivering O(log n) lookups with less complexity than B-trees. This post explains how they work, why randomness is useful and why you might choose them over B-trees.
LLM prompt to get a useful summary based on 5 questions
Data modeling seems straightforward—until you realize the real world is messy, vague, and full of philosophical dilemmas. In this post, I dive into the challenges of defining "one thing," handling change, and categorizing entities in databases. Spoiler: It’s way harder than it looks.
How effective are LLMs as tutors? There are some crazy anecdotes out there about how they can replace teachers. But this study is one of the first I've seen that actually tries to understand its effect, _and_ what happens after it's taken away.
When building a service that handles many different API integrations, it's helpful to standardize the response. The `Result` monad is useful for expressing the response as either `Success` (everything went well), or `Failure` (something has gone wrong).
Memory Management for SimpleDB. Implementing the buffer and buffer manager.
A practical guide to Python concurrency using a Fibonacci server to demonstrate the strengths and limitations of threads, processes, and asyncio. Learn when to use each approach through hands-on examples
disk and file management
RAG is not all you need. This post will cover some of the common problems that are encountered in a simple RAG system, and potential solutions for them.
Interpolating vector and lexical search gives better results than either alone
As a fellow grug brained developer, I also agree that complexity is very bad. But what is software complexity? What does it look like? How does it come about?
I've been working my way through Andrej Karpathy's 'spelled-out intro to backpropagation', and this post is my recap of how backpropagation works.
Step-by-step walkthrough of using GPT for debugging.
Stable diffusion allows non-artistic individuals, like myself, to create stunning images simply by providing a text prompt.
b-trees • database indexes • partial index • function-based index • multi-column index
grouping similar data points into sets • clustering methods • kmeans
Tables with polymorphic associations can be quite annoying to query. Use views to make it much easier.
Steps that will make future investigation of your Postgres performance much easier.