hackertube.io

State of GPT | BRK216HFS

Microsoft Developer · talk · Watch on YouTube

LLMs & generative AIMachine learning & deep learning

What it is

Andrej Karpathy walks through how GPT assistants are trained, from tokenization and pretraining to supervised finetuning and RLHF, and gives practical prompting strategies and a view of the tooling ecosystem.

Why Hacker News recommends it

Recommended as the clearest overview of how LLM assistants are built and how to use them well.

  • 6 people recommended it
  • 3 different months
  • 1 strong recommendations
  • first 2023, last 2025

What HN commenters said

  1. “It changed the way I use GPT-4.”

    Explains how LLMs work with good prompting ideas

    sagarpatil, 2023-06 · Strong recommendation · in “GPT Best Practices”

  2. “provides a through introduction to the process used to train”

    Deep summary of GPT training phases and techniques

    mcint, 2023-06 · Recommendation · in “It’s infuriatingly hard to understand how closed models train on their input”

  3. “(~hr) Karpathy talk”

    Hour-long Karpathy talk listed among good prompting guides

    valgaze, 2023-06 · Recommendation · in “A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT”

  4. “Karpathy explains it far better than I did in his”

    Karpathy explains LLM reasoning and prompting well

    jsight, 2023-06 · Recommendation · in “The future of AI according to thousands of forecasters”

  5. “State of GPT by Karpathy”

    Listed among resources for getting up to speed on LLMs

    tikkun, 2024-11 · Recommendation · in “Ask HN: Recommendation for a SWE looking to get up to speed with latest on AI”

  6. “Explains many details in a form suitable for beginners.”

    Beginner-friendly explanation of the stages of LLM training

    tintor, 2025-02 · Recommendation · in “RLHF Book”

Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.

Topics: llm, rlhf, prompting, karpathy, fine-tuning

Also recommended by the same people