Why e is special

What you’ll learn

You have met e ≈ 2.71828 and been told it is important, probably without being told why. The single capability this lesson builds: explain why e^x is its own derivative, and connect that property to exponential (self-proportional) growth, so e stops being an arbitrary decimal and becomes inevitable.

The defining behavior is d/dx(e^x) = e^x: the rate of change equals the value at every point. You will see that the derivative of any exponential is d/dx(a^x) = M(a)·a^x, the function itself times a base-dependent constant; that this multiplier is below 1 at base 2 (≈ 0.693) and above 1 at base 3 (≈ 1.099), so it crosses exactly 1 at some base in between, and that base is e. You will check the property in numbers, combine it with the chain rule to get d/dx(e^(kx)) = k·e^(kx), recognize that as the solution to f'(x) = k·f(x) (“rate proportional to current value”), and see why that makes e the natural shape of compound interest, population growth, and radioactive decay. A worked sigmoid derivative (σ' = σ(1-σ)) shows the same property at work inside machine learning.

Where this fits

This is lesson 7 of Phase 2 (The differentiation toolkit). It leans directly on the chain rule (lesson 6) to handle e^(kx), and it sets up two later lessons: the Taylor series of e^x (lesson 13) follows from the self-derivative property, and e is the calculus prerequisite for softmax and the sigmoid in the neural-network tracks (Track 11). Phase 2 continues with implicit differentiation (lesson 8) and limits (lesson 9).

Before you start

Prerequisite (within this track): lesson 6, The chain rule, since d/dx(e^(kx)) = k·e^(kx) is the self-derivative property combined with the chain rule, and the sigmoid example chains through e^(-x). It helps to be comfortable with exponent rules (especially a^(x+h) = a^x · a^h). No coding, nothing installed; the practice is pen and paper with a calculator for the numeric check.

By the end, you’ll be able to

• Explain that e is defined as the unique base for which e^x is its own derivative, not by its digits
• Show that the derivative of any exponential a^x is M(a)*a^x, and why the multiplier equals 1 at base e
• Use the chain rule to differentiate e^(kx) = ke^(kx) and recognize it as the solution to f’ = kf
• Connect e to self-proportional change (growth, decay) and to softmax and the sigmoid in machine learning

Time and difficulty

• Read time: about 11 minutes
• Practice time: about 13 minutes (a numeric self-derivative check, differentiating exponentials, a rate-proportional-to-value verification, and flashcards)
• Difficulty: standard