Loading video...
Video Failed to Load
Hugo Larochelle explains how a neuron computes its output: first, it calculates the preactivation, the sum of the bias and weighted inputs, and then applies an activation function.
29,307 views • 11 months ago •via X (Twitter)
24 Comments

👌

Macrohard

@grok can neurons compute?

Note: "artificial" neuron, the legacy of the perceptron as adapted for neural networks (AI). I've been thinking about this: There's no way biologic neuronal architecture is serialized like this (too slow), so an updated NN would be parallel (system of equations or quantum)

Trash explanation

👏👏👏🎯

Hugo Larochelle's neuron explanation? Yawn another ivory-tower egghead regurgitating basic neural net math like it’s divine revelation. Preactivation, weights, bias, activation function? This is undergrad AI 101, not groundbreaking insight. While he’s preaching to the choir, Big Tech’s black-box algorithms built on this oversimplified drivel are fleecing privacy, rigging elections, and automating job theft. Larochelle’s not enlightening us; he’s shilling for the AI overlords, dressing up their dystopian code as "science." Wake up, sheep this neuron worship is just propaganda for the singularity scam!

We are amazing computers

👌

Neurons out here doing math while we struggle to remember our PINs. 😅🧠

Pure Maths 😎 the oldest Science

Hugo Larochelle breaks it down simply — preactivation plus activation function equals neuron output. 👀✨

Elegant

Neurons: “We sum, we activate, we conquer.” Humans: still can’t calculate a tip. 😏💡

Lol. It's complicated as hell, but it's fun to learn.

So simple at the heart of it.

3blue1brown has some very beautiful videos on this god I wish I were 18 again so I could study this stuff instead of the boring crap I do

Thanks for sharing. Planning on making a neural network soon?

@grok explain better

~ thanks for this clip slatt!

@grok what is being explained. I’m lost. Explain it to me like I missed this class

I remember when I was studying at the university about this in a AI course at 2010, we make a neuron with OpAmps (electronics) and We was entry a lot of signals for watch your results. Today the AI is an reality!

@grok ¿Qué tipo de persona entiende esto, un matemático o un programador?

When a neuron casually does complex math every millisecond and you’re like… 💀🤯

