Skip to content
tezvyn:

All bites

The whole library, newest first. Filter by what you are here for, or pick a topic if you already know.

8667 bites

Page 434

Word Embeddings: Turning Words into Vectors
LLMs & Generative AI2 min read

Word Embeddings: Turning Words into Vectors

Word embeddings turn words into numerical vectors, like coordinates on a map of meaning. Words with similar meanings, like "king" and "queen," are placed close together in this vector space. This is fundamental for text analysis in machine learning, allowing models to grasp semantic relationships instead of just matching text. The footgun is assuming the vector's individual numbers are human-interpretable; they are abstract features learned from data.

LLMs & Generative AI2 min read

Byte Pair Encoding: Compressing Text for LLMs

Think of Byte Pair Encoding (BPE) as creating custom abbreviations for common letter pairs to compress text. It repeatedly finds the most frequent pair, like 'th', and merges it into a new token. LLMs use this to build vocabularies of common sub-word units, helping them understand rare words. The main footgun is that the final vocabulary size is fixed; choosing the wrong size can hurt model performance and efficiency.

LLMs & Generative AI2 min read

Large Language Models (LLMs)

A large language model is a sophisticated pattern-matching engine trained on a massive library of text. They power modern chatbots and can generate, summarize, or translate text by predicting the most probable next word based on the patterns they've learned. The key footgun is that their output reflects the biases and inaccuracies of their training data, making them confident but potentially unreliable.

Top 10 FastAPI interview questions for senior roles
Python & FastAPI2 min read

Top 10 FastAPI interview questions for senior roles

FastAPI interview questions rarely test syntax. They probe whether you understand its three pillars: ASGI powered async concurrency, Pydantic validation at the request boundary, and the Depends dependency injection graph that wires services together.

Async SQLAlchemy 2.0: the mental model that clicks
Python & FastAPI2 min read

Async SQLAlchemy 2.0: the mental model that clicks

SQLAlchemy's async layer bridges its synchronous ORM internals to asyncio. Use AsyncSession and AsyncEngine with an async driver, and await database work. Prerequisites: Python asyncio and basic SQLAlchemy ORM; lazy relationships must be loaded explicitly to avoid implicit I/O errors.

Server Components, explained without the jargon
React & Next.js2 min read

Server Components, explained without the jargon

React Server Components render on the server and send the client a description of the result, not JavaScript. That means no bundle cost for server rendered parts and direct database access, with client components handling only the interactive pieces.

React interview cheatsheet: hooks rules in 60 seconds
React & Next.js2 min read

React interview cheatsheet: hooks rules in 60 seconds

The rules of hooks exist because React identifies each hook by call order, not by name. Call hooks only at the top level of a component or custom hook, never inside a loop, condition, or nested function, or state attaches to the wrong hook.