But how do AI images and videos actually work? | Guest video by Welch Labs
3Blue1Brown1 HN recommender · recommended in 1 month, 2025-2026
2.2M views · 1 year ago · explainer · evergreen
A guest video by Welch Labs on 3Blue1Brown explaining how text-to-image and video models work. It covers CLIP, shared embedding spaces, diffusion models (DDPM, DDIM), conditioning, guidance and negative prompts.
HN uses it as a clear explainer for diffusion mechanics, such as negative prompts and faster sampling, and for joint text-image encoders.
27:14Transformers, the tech behind LLMs | Deep Learning Chapter 5
3Blue1Brown
11.6M views • 2 years ago
26:10Attention in transformers, step-by-step | Deep Learning Chapter 6
3Blue1Brown
4.8M views • 2 years ago
8:25Large Language Models from scratch
Graphics in 5 Minutes
375K views • 4 years ago
28:48LSTM is dead. Long Live Transformers!
Seattle Applied Deep Learning
540K views • 6 years ago
3:31:24Deep Dive into LLMs like ChatGPT
Andrej Karpathy
10.2M views • 1 year ago
![Neural Networks Demystified [Part 1: Data and Architecture]](https://i.ytimg.com/vi/bxe2T-V8XRs/hq2.jpg)
3:07Neural Networks Demystified [Part 1: Data and Architecture]
Welch Labs
838K views • 12 years ago
20:39AI Language Models & Transformers - Computerphile
Computerphile
354K views • 7 years ago
1:26Fake Obama created using AI video tool - BBC News
BBC News
1.7M views • 9 years ago
1:56:20Let's build GPT: from scratch, in code, spelled out.
Andrej Karpathy
8.2M views • 3 years ago
23:02Did We Just Change Animation Forever?
Corridor Crew
3.5M views • 3 years ago
27:07Attention Is All You Need
Yannic Kilcher
814K views • 8 years ago
1:22:38CS480/680 Lecture 19: Attention and Transformer Networks
Pascal Poupart
374K views • 7 years ago
1:31:13A Hackers' Guide to Language Models
Jeremy Howard
580K views • 3 years ago
6:18A Style-Based Generator Architecture for Generative Adversarial Networks
Tero Karras FI
1.3M views • 7 years ago
42:40State of GPT | BRK216HFS
Microsoft Developer
785K views • 3 years ago
57:57Lecture 1 | Introduction to Convolutional Neural Networks for Visual Recognition
Stanford University School of Engineering
2.9M views • 9 years ago
4:17AI Learns Noise Filtering For Photorealistic Videos | Two Minute Papers #215
Two Minute Papers
65K views • 8 years ago
24:07AI can't cross this line and we don't know why.
Welch Labs
2.5M views • 2 years ago
39:43Do we need Attention? - Linear RNNs and State Space Models (SSMs) for NLP
Sasha Rush
35K views • 3 years ago
13:52Ch(e)at GPT? - Computerphile
Computerphile
617K views • 3 years ago
3 comments from Hacker News
> Because you somehow need a giant training set which describes images in natural language, no? That's definitely one way - they train a text encoder together with an image encoder on a labelled set of images. WL & 3b1b made a nice video on it: ▶ this video
Breakthrough in image generation speed literally came from applying better differential equations for diffusion taken from statistical mechanics physics papers: ▶ this video
My favourite instantiation of this weird invocation is from this AI video generator, where they literally subtract the prompt for 'low quality video' from the input, and it improves the quality. ▶ this video I've just migrated my AI product to a different underlying model and had to redo a few of the prompts that the new model was interpreting differently. It's not obseleted, just requires a bit of migration. The improved quality of the new models outweighs any issues around prompting.