Skip to content

Commit c0a77fa

Browse files
committed
upd. page. math - derivative.
1 parent c9f21c4 commit c0a77fa

File tree

253 files changed

+24221
-1400
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

253 files changed

+24221
-1400
lines changed

blog/content/page/playlists/index.en.md

Lines changed: 6 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ menu:
2020
## 👥 On Their Shoulders
2121

2222
- **[The Future of AI: Why Scaling Alone is No Longer Enough](/en/p/ai-future-beyond-scaling/)**
23-
- 🎬 [Video](https://youtu.be/5mUEOx3uLDc) · reading 12 min / video 8 min
23+
- 🎬 [Video](https://youtu.be/5mUEOx3uLDc) · reading 12 min / video 8 min (original 52 min)
2424
- 👥 Participants: Nicholas Thompson, Eric Xing, Yoshua Bengio, Yuval Noah Harari, Yejin Choi
2525
- 🏷️ Terms: scaling, reward hacking, world model, guardrail, AGI, anthropomorphism
2626
- 📊 Difficulty: intermediate
@@ -32,7 +32,11 @@ menu:
3232
- 📋 Overview: which math areas are needed for AI/ML and why
3333
- 🏷️ Linear algebra, calculus, probability and statistics
3434
- 📊 Difficulty: basic
35+
- **[Math Analysis — Example of Analysis (Finding πr²)](/en/p/math-circle-area/)**
36+
- 📋 Slicing the circle into rings, unrolling into strips 2πr × Δr — yields a triangle of area πR²
37+
- 🏷️ Integral, limits, area of circle
38+
- 📊 Difficulty: basic
3539
- **[Math Analysis — Derivatives](/en/p/math-derivatives/)**
36-
- 📋 Derivative, gradient, chain rule — backbone of neural network training
40+
- 📋 Without this you can't understand how neural networks learn. Derivative, gradient, chain rule.
3741
- 🏷️ Gradient descent, backpropagation
3842
- 📊 Difficulty: basic

blog/content/page/playlists/index.md

Lines changed: 7 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ menu:
2020
## 👥 На их плечах
2121

2222
- **[Будущее ИИ: почему одного масштабирования уже недостаточно](/p/ai-future-beyond-scaling/)**
23-
- 🎬 [Видео](https://youtu.be/5mUEOx3uLDc) · чтение 12 мин / видео 8 мин
23+
- 🎬 [Видео](https://youtu.be/5mUEOx3uLDc) · чтение 12 мин / видео 8 мин (оригинал 52 мин)
2424
- 👥 Участники: Nicholas Thompson, Eric Xing, Yoshua Bengio, Yuval Noah Harari, Yejin Choi
2525
- 🏷️ Термины: масштабирование, reward hacking, world model, guardrail, AGI, антропоморфизм
2626
- 📊 Сложность: средняя
@@ -32,7 +32,11 @@ menu:
3232
- 📋 Обзор: какие разделы математики нужны для ИИ/МО и зачем
3333
- 🏷️ Линейная алгебра, анализ, вероятность и статистика
3434
- 📊 Сложность: базовая
35-
- **[Мат Анализ — Производные](/p/math-derivatives/)**
36-
- 📋 Производная, градиент, правило цепочки — основа обучения нейросетей
35+
- **[Мат анализ — пример анализа (нахождение πr²)](/p/math-circle-area/)**
36+
- 📋 Нарезка круга на кольца, разворот в полоски 2πr × Δr — получается треугольник с площадью πR²
37+
- 🏷️ Интеграл, пределы, площадь круга
38+
- 📊 Сложность: базовая
39+
- **[Мат анализ — Производные](/p/math-derivatives/)**
40+
- 📋 Без этого не понять, как нейросеть вообще учится. Производная, градиент, правило цепочки.
3741
- 🏷️ Градиентный спуск, backpropagation
3842
- 📊 Сложность: базовая
453 KB
Loading

blog/content/post/ai-basics-intro/index.en.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
---
22
title: "AI basics – introduction"
33
description: "A short introduction to the fundamentals of artificial intelligence"
4-
date: "2025-03-03"
4+
date: "2026-03-03"
55
slug: "ai-basics-intro"
66
tags:
77
- Artificial Intelligence

blog/content/post/ai-basics-intro/index.md

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,12 +1,13 @@
11
---
22
title: "Базовый минимум про ИИ – введение"
33
description: "Краткое введение в основы искусственного интеллекта"
4-
date: "2025-03-03"
4+
date: "2026-03-03"
55
slug: "ai-basics-intro"
66
tags:
77
- Искусственный интеллект
88
- Машинное обучение
99
- База
10+
image: cover.jpg
1011
---
1112

1213
Это первая статья из цикла «Базовый минимум» — коротко про то, как устроен ИИ. Каждая статья будет освещать одно понятие или концепцию, постараюсь делать их короткими и последовательными.
242 KB
Loading

blog/content/post/ai-future-beyond-scaling/index.en.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
---
22
title: "The Future of AI: Why Scaling Alone is No Longer Enough"
33
description: "Scaling brought us LLMs, but the next leap requires a paradigm shift in how we train models"
4-
date: "2025-03-10"
4+
date: "2026-03-10"
55
slug: "ai-future-beyond-scaling"
66
tags:
77
- Artificial Intelligence
@@ -10,7 +10,7 @@ tags:
1010

1111
Until recently, the main recipe for progress in AI seemed obvious: more data, more compute, more parameters. And it worked — it led to modern large language models. But the central idea is that **scaling is useful but no longer sufficient**. The next step requires not just more power, but a paradigm shift in how we train models.
1212

13-
**Article based on a video.** [Watch on YouTube](https://youtu.be/5mUEOx3uLDc)
13+
**Article based on a video.** [Watch on YouTube](https://youtu.be/5mUEOx3uLDc) (8 min) · [Original 52 min](https://youtu.be/MdGnCIl-_hU?si=ylocQLiEBvp9Bydb)
1414

1515
<details>
1616
<summary>Panel participants and their backgrounds</summary>

blog/content/post/ai-future-beyond-scaling/index.md

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,16 +1,17 @@
11
---
22
title: "Будущее ИИ: почему одного масштабирования уже недостаточно"
33
description: "Масштабирование привело к LLM, но следующий шаг потребует смены парадигмы обучения"
4-
date: "2025-03-10"
4+
date: "2026-03-10"
55
slug: "ai-future-beyond-scaling"
66
tags:
77
- Искусственный интеллект
88
- Машинное обучение
9+
image: cover.jpg
910
---
1011

1112
Ещё недавно главный рецепт прогресса в искусственном интеллекте казался очевидным: больше данных, больше вычислений, больше параметров. И этот рецепт действительно сработал — именно он привёл к появлению современных больших языковых моделей. Но центральная идея в том, что **масштабирование полезно, но уже недостаточно**. Следующий шаг потребует не просто наращивания мощностей, а смены парадигмы обучения.
1213

13-
**Статья по видео.** [Смотреть на YouTube](https://youtu.be/5mUEOx3uLDc)
14+
**Статья по видео.** [Смотреть на YouTube](https://youtu.be/5mUEOx3uLDc) (8 мин) · [Оригинал 52 мин](https://youtu.be/MdGnCIl-_hU?si=ylocQLiEBvp9Bydb)
1415

1516
<details>
1617
<summary>Кто участники панели и чем они известны</summary>
247 KB
Loading
Lines changed: 76 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,76 @@
1+
---
2+
title: "Math Analysis — Example of Analysis (Finding πr²)"
3+
description: "Slicing the circle into rings, unrolling into strips — and area πR² via a triangle"
4+
date: "2026-03-12"
5+
slug: "math-circle-area"
6+
tags:
7+
- Machine Learning
8+
- Mathematics
9+
---
10+
11+
How do we find the area of a circle **analytically**?
12+
13+
## Slicing into Rectangles
14+
15+
**Idea:** slice the circle into thin concentric rings and «unroll» each ring into a rectangle.
16+
17+
A ring of radius r and width Δr has circumference 2πr. Unrolled, it becomes a rectangle:
18+
- length: 2πr
19+
- width: Δr
20+
- area: 2πr · Δr
21+
22+
The total area is the sum of all such rings, from r = 0 to r = R. The smaller Δr, the better the approximation.
23+
24+
<svg viewBox="0 0 360 200" xmlns="http://www.w3.org/2000/svg" style="max-width: 360px; display: block; margin: 1em 0;">
25+
<defs>
26+
<linearGradient id="ring-en" x1="0%" y1="0%" x2="100%" y2="0%">
27+
<stop offset="0%" stop-color="#3b82f6"/>
28+
<stop offset="100%" stop-color="#60a5fa"/>
29+
</linearGradient>
30+
</defs>
31+
<circle cx="100" cy="100" r="85" fill="none" stroke="#4b5563" stroke-width="1"/>
32+
<path d="M 100 15 A 85 85 0 0 1 185 100 A 85 85 0 0 1 100 185 A 85 85 0 0 1 15 100 A 85 85 0 0 1 100 15" fill="url(#ring-en)" fill-opacity="0.12" stroke="#60a5fa" stroke-width="1.5"/>
33+
<path d="M 100 30 A 70 70 0 0 1 170 100 A 70 70 0 0 1 100 170 A 70 70 0 0 1 30 100 A 70 70 0 0 1 100 30" fill="url(#ring-en)" fill-opacity="0.2" stroke="#60a5fa" stroke-width="1.5"/>
34+
<path d="M 100 45 A 55 55 0 0 1 155 100 A 55 55 0 0 1 100 155 A 55 55 0 0 1 45 100 A 55 55 0 0 1 100 45" fill="url(#ring-en)" fill-opacity="0.28" stroke="#60a5fa" stroke-width="1.5"/>
35+
<path d="M 100 60 A 40 40 0 0 1 140 100 A 40 40 0 0 1 100 140 A 40 40 0 0 1 60 100 A 40 40 0 0 1 100 60" fill="url(#ring-en)" fill-opacity="0.35" stroke="#60a5fa" stroke-width="1.5"/>
36+
<path d="M 100 75 A 25 25 0 0 1 125 100 A 25 25 0 0 1 100 125 A 25 25 0 0 1 75 100 A 25 25 0 0 1 100 75" fill="url(#ring-en)" fill-opacity="0.5" stroke="#60a5fa" stroke-width="1.5"/>
37+
<circle cx="100" cy="100" r="12" fill="#1e3a5f" stroke="#3b82f6" stroke-width="2"/>
38+
<line x1="100" y1="100" x2="100" y2="30" stroke="#f59e0b" stroke-width="2" stroke-dasharray="4 2"/>
39+
<text x="115" y="60" font-size="14" fill="#f59e0b">R</text>
40+
<text x="100" y="108" font-size="12" fill="#e5e7eb" text-anchor="middle">r=0</text>
41+
<text x="250" y="100" font-size="13" fill="#9ca3af">Rings r = 0 … R</text>
42+
</svg>
43+
44+
If we arrange these cut strips (each of length 2πr) on a graph from left to right — from r = 0 to r = R — we get a triangle. Base R, height 2πR: the area of the triangle is ½ · R · 2πR = πR².
45+
46+
<svg viewBox="0 0 320 220" xmlns="http://www.w3.org/2000/svg" style="max-width: 320px; display: block; margin: 1em 0;">
47+
<defs>
48+
<linearGradient id="strip-en" x1="0%" y1="0%" x2="100%" y2="0%">
49+
<stop offset="0%" stop-color="#3b82f6"/>
50+
<stop offset="100%" stop-color="#60a5fa"/>
51+
</linearGradient>
52+
</defs>
53+
<!-- Axes -->
54+
<line x1="40" y1="180" x2="260" y2="180" stroke="#6b7280" stroke-width="1.5"/>
55+
<line x1="40" y1="180" x2="40" y2="20" stroke="#6b7280" stroke-width="1.5"/>
56+
<!-- Strips: height 2πr at left edge (not above the line), width Δr -->
57+
<rect x="40" y="172" width="22" height="8" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
58+
<rect x="62" y="164" width="22" height="16" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
59+
<rect x="84" y="148" width="22" height="32" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
60+
<rect x="106" y="132" width="22" height="48" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
61+
<rect x="128" y="116" width="22" height="64" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
62+
<rect x="150" y="100" width="22" height="80" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
63+
<rect x="172" y="84" width="22" height="96" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
64+
<rect x="194" y="68" width="22" height="112" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
65+
<rect x="216" y="52" width="22" height="128" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
66+
<rect x="238" y="36" width="22" height="144" fill="url(#strip-en)" fill-opacity="0.6" stroke="#60a5fa" stroke-width="1"/>
67+
<!-- Line 2πr (from r=0 to r=R) -->
68+
<line x1="40" y1="180" x2="260" y2="20" stroke="#f59e0b" stroke-width="2"/>
69+
<text x="270" y="185" font-size="14" fill="#9ca3af">r</text>
70+
<text x="25" y="35" font-size="14" fill="#9ca3af" text-anchor="end">2πr</text>
71+
<text x="255" y="195" font-size="12" fill="#f59e0b">R</text>
72+
<text x="35" y="45" font-size="12" fill="#f59e0b" text-anchor="end">2πR</text>
73+
<text x="150" y="110" font-size="12" fill="#9ca3af" text-anchor="middle">½·R·2πR = πR²</text>
74+
</svg>
75+
76+
**Result:** area of the circle S = πR². The same approach — slicing a shape into «small pieces», summing their areas, and taking the limit — underlies integral calculus and, as we will see, leads to derivatives.

0 commit comments

Comments
 (0)