HackerTube.dev

Large Language Models from scratch

Graphics in 5 Minutes · 375.4K views · 2022-07-17 · 8:25

Open on YouTube ↗

If playback is restricted in your region, try opening the original video. Use Mark watched to add it to your marked videos.

Overview

A short cartoon explaining how language models like ChatGPT work, from autocomplete and conditional probabilities to neural networks, gradient descent and backpropagation. It shows how networks act as universal function approximators.

On Hacker News

HN recommends it as a quick, non-dumbed-down way to get up to speed on how LLMs are built.

Recommended by 3 HN people across 2 months · HN mentions 2023–2024

Topics

LLMs & generative AIMachine learning & deep learning
Search by tags
llmneural networksgradient descenttransformerslanguage modeling

What people said

peter_d_sherman · 2023-05 · Strong recommendation

Consider the following frame/time from the video: [VIDEO] At this point in time in the video, we see the following 3 things: 1) Fourier Series 2) Taylor Series 3) Neural Network The remainder of the video goes on to show how Neural Networks can approximate functions -- much like the Fourier Series and Taylor Series can... Now, I don't hold an advanced Ph.D. degree in Math, and prior to watching this video, I had never made the connection between Fourier Series, Taylor Series, and Neural Networks... I knew about each of those independently -- but prior to watching the video, I never made the connection that they can be used to approximate one another. And that's the one of the key words -- approximate . The other key word is Identity -- as in Mathematical Identity ... That's because IF two Mathematical Objects (functions, algorithms, call them what-you-will) are approximately equal -- then there should be a set of conditions (i.e., "If and only if...", etc.), a set of rigorously defined mathematical constraints which, if those conditions/constraints are met, then the two Mathematical Objects are exactly precisely equivalent -- under those rigorously defined conditions/constraints.... In other words, under those rigorously defined conditions/constraints -- one Mathematical Object could be substituted for another -- with no loss of precision or result. In other words, under those rigorously defined conditions/constraints -- those two completely apparently different Mathematical

Read the original HN comment ↗
sainez · 2023-04 · Strong recommendation

If anyone is looking for a quick overview of how LLMs are built, I highly recommend this video by Steve Seitz: [VIDEO] . It does an excellent job of taking you from 0 to a decent understanding without dumbing down the content or abusing analogies.

Read the original HN comment ↗
runetech · 2023-05 · Recommendation

Not sure I can give an ELI5 explanation, but for me this the closest to what you are after in two short videos on the topic. At least they helped me get up to speed fast. Hope they will do the same for you ;-) Large Language Models from scratch [VIDEO] Large Language Models: Part 2 [LINK]

Read the original HN comment ↗

Sources

Video by Graphics in 5 Minutes, played through YouTube.

Data and selection notes

HN summaries and labels may contain AI errors. HN dates refer to discussions, not video publication.

HN Library: HackerTube export (2026-10-31) and open-index/hacker-news. Upstream summaries and classifications may contain AI errors.

Views, duration and playback status refresh separately from the published catalog. When an update fails, existing data remains visible.

Selection checks
  • SaaS & AI candidates: Matches the configured criteria
  • Editor-approved picks: Editorial review is required

Rule matches identify candidates. They do not guarantee video quality or financial results.

About the library and its sources