Étagère mesuréeIA et apprentissage automatique
AI and machine learning videos, ranked by how much of them is useful
- Remplissage médian
- 37%
- La partie utile commence
- 1:08
- Durée typique
- 24 min
dernier recalcul 16 sept. 2026
This is the category where the gap between reputation and substance is widest, because the subject rewards two very different videos that look identical from the outside.
The first kind spends its runtime on analogy. Neurons are like brain cells, attention is like paying attention, the model is like a very good autocomplete. It is pleasant, it is often accurate, and after forty minutes you cannot do anything you could not do before. The second kind uses one analogy to get you onto the ramp and then shows the actual arithmetic. Both get the same thumbnail treatment, and both are described as “explained”.
The measurement that separates them is density: what share of the runtime introduces something the previous minute did not already contain. Repetition of an idea in three different metaphors scores low here, and that is deliberate — it is the single most common way a fifteen-minute idea becomes a fifty-minute video in this subject.
Entries also carry a difficulty level, because “explained simply” and “explained” are not the same shelf, and the fastest way to waste an hour is to pick the wrong one for where you already are.
-
01
Let's reproduce GPT-2 (124M)
0:004:01:26- Densité
- 84
- Remplissage
- 27%
- Utile à partir de
- 3:34
A code-complete, first-principles GPT-2 reproduction — one of the most valuable hands-on deep learning tutorials available.
-
02
But what is a convolution?
0:0023:01- Densité
- 84
- Remplissage
- 24%
- Utile à partir de
- 1:41
A masterclass build-up of convolution — from dice probabilities to a genuinely surprising O(n log n) FFT algorithm.
-
03
Attention in transformers, step-by-step | Deep Learning Chapter 6
0:0026:10- Densité
- 83
- Remplissage
- 21%
- Utile à partir de
- 1:40
A masterfully clear, dense walkthrough of the attention mechanism that rewards careful watching with real technical understanding.
-
04
But what is a neural network? | Deep learning chapter 1
0:0018:40- Densité
- 81
- Remplissage
- 26%
- Utile à partir de
- 2:39
Une explication rare, véritablement construite depuis zéro, rigoureuse et re-démontrable, de ce qu'est réellement la mathématique d'un réseau de neurones.
-
05
Transformers, the tech behind LLMs | Deep Learning Chapter 5
0:0027:14- Densité
- 80
- Remplissage
- 26%
- Utile à partir de
- 1:26
A masterclass primer on transformer internals — dense, rigorous, and exactly what its title promises.
-
06
Deep Dive into LLMs like ChatGPT
0:003:31:24- Densité
- 82
- Remplissage
- 27%
- Utile à partir de
- 1:07
One of the clearest, most complete plain-English explanations available of how ChatGPT-style LLMs are actually built and why they behave the way they do.
-
07
Let's build GPT: from scratch, in code, spelled out.
0:001:56:20- Densité
- 82
- Remplissage
- 35%
- Utile à partir de
- 14:11
A masterclass build-along: real working GPT code, the actual mechanics behind ChatGPT, with almost no filler.
-
08
But how do AI images and videos actually work? | Guest video by Welch Labs
0:0037:20- Densité
- 82
- Remplissage
- 22%
- Utile à partir de
- 3:28
A dense, physics-grounded explainer that actually teaches how diffusion models and CLIP combine to generate AI images and video — rare, high-value technical content.
-
09
But what is quantum computing? (Grover's Algorithm)
0:0036:54- Densité
- 82
- Remplissage
- 24%
- Utile à partir de
- 0:52
A rigorous, honest deep dive that builds Grover's algorithm from first principles — 3Blue1Brown at its best.
-
10
What is backpropagation really doing? | Deep learning chapter 3
0:0012:47- Densité
- 80
- Remplissage
- 33%
- Utile à partir de
- 0:53
Une masterclass pour construire une intuition authentique de la rétropropagation sans une seule formule — un enseignement de tout premier ordre.
-
11
[1hr Talk] Intro to Large Language Models
0:0059:48- Densité
- 80
- Remplissage
- 25%
- Utile à partir de
- 3:31
Une explication dense, sans publicité, de niveau masterclass, sur la façon dont les LLM sont construits, utilisés et attaqués — parmi les meilleures vidéos éducatives générales sur l'IA…
-
12
Backpropagation calculus | Deep Learning Chapter 4
0:0010:18- Densité
- 80
- Remplissage
- 25%
- Utile à partir de
- 0:32
A tight, rigorous derivation of backprop's chain-rule math — dense, durable, exactly what the title promises.
-
13
Gradient descent, how neural networks learn | Deep Learning Chapter 2
0:0020:33- Densité
- 79
- Remplissage
- 28%
- Utile à partir de
- 1:52
Une masterclass pour construire une intuition authentique de la descente de gradient, honnête sur les véritables limites de la méthode.
-
14
Transformer Neural Networks, ChatGPT's foundation, Clearly Explained!!!
0:0036:15- Densité
- 79
- Remplissage
- 23%
- Utile à partir de
- 1:20
A rigorous, worked-numbers walkthrough of transformer internals that actually teaches how ChatGPT-style models work, not just what they do.
-
15
Let's build the GPT Tokenizer
0:002:13:35- Densité
- 82
- Remplissage
- 31%
- Utile à partir de
- 4:22
An exceptionally dense, hands-on build of a real GPT tokenizer that also explains — with concrete transcript-verifiable examples — why tokenization is the hidden cause of most weird LLM…
-
16
How might LLMs store facts | Deep Learning Chapter 7
0:0022:43- Densité
- 82
- Remplissage
- 22%
- Utile à partir de
- 0:58
A rigorous, self-contained walkthrough of how transformer MLP blocks might encode facts, capped with a genuinely illuminating dive into superposition and near-orthogonality in high…
-
17
The Most Important Algorithm in Machine Learning
0:0040:08- Densité
- 78
- Remplissage
- 27%
- Utile à partir de
- 1:28
A rigorous, ground-up derivation of backpropagation that earns its title without needing hype.
-
18
Why Does Diffusion Work Better than Auto-Regression?
0:0020:18- Densité
- 79
- Remplissage
- 32%
- Utile à partir de
- 0:32
A rigorous, original explanation of why diffusion models generate images faster than autoregressive ones — genuinely worth the watch.
-
19
The Essential Main Ideas of Neural Networks
0:0018:54- Densité
- 76
- Remplissage
- 29%
- Utile à partir de
- 1:54
A rare, fully worked walkthrough of the actual math inside a neural network — StatQuest at its clearest.
-
20
Deep RL Bootcamp Lecture 4B Policy Gradients Revisited
0:0034:55- Densité
- 76
- Remplissage
- 28%
- Utile à partir de
- 1:08
A superb, intuitive deep dive into policy gradients with a real code walkthrough — genuinely teaches, nothing being sold.
-
21
Flow-Matching vs Diffusion Models explained side by side
0:0016:08- Densité
- 75
- Remplissage
- 32%
- Utile à partir de
- 0:33
A dense, well-structured technical breakdown that delivers exactly what its title promises: diffusion vs flow matching, math and all.
-
22
Diffusion Models | DDPM Explained
0:0029:29- Densité
- 81
- Remplissage
- 20%
- Utile à partir de
- 1:18
A rigorous, self-derived walkthrough of DDPM math that earns its 'explained' title with real depth, not just paper restatement.
-
23
Diffusion Models: DDPM | Generative AI Animated
0:0032:06- Densité
- 79
- Remplissage
- 22%
- Utile à partir de
- 1:37
A dense but genuinely rigorous derivation of DDPM from theory to working code — one of the clearer deep explainers on the topic.
-
24
Diffusion Models From Scratch | Score-Based Generative Models Explained | Math Explained
0:0038:11- Densité
- 79
- Remplissage
- 27%
- Utile à partir de
- 1:20
A rigorous, derivation-heavy walkthrough that genuinely explains where the diffusion model equations come from, not just what they are.
-
25
The AI Progress Chart Everyone Is Misreading — Beth Barnes & David Rein
0:001:53:27- Densité
- 78
- Remplissage
- 28%
- Utile à partir de
- 3:58
A rare, unusually candid deep-dive by the researchers themselves into how the AI 'progress chart' is built, where it breaks, and why the public keeps over-reading it.
Gistil's own measurement. Not a YouTube rating, and not the channel's position. Le classement est le nôtre ; les vidéos appartiennent à leurs chaînes.
Comment ce classement est établi
Chaque vidéo ici a été mesurée sur les mêmes axes — quelle part de sa durée porte de l'information, quelle part est du remplissage, et à quelle seconde elle commence à payer. L'ordre suit la densité de valeur, pas les vues, la fraîcheur ou à quel point on a aimé la chaîne. Ce que signifient les chiffres →