How AI works, Explained with Mushrooms ! (Easiest Explanation )
Xplained with MUSHROOMS
0:00 / 0:00
How AI works, Explained with Mushrooms ! (Easiest Explanation )
56 838 просмотров · 4 дня назад
Xplained with MUSHROOMS
363 подписчика
56 838 просмотров · 4 дня назад
How ChatGPT and other AI chatbots actually work, explained with mushrooms: training, weights, tokens, how an answer is built one piece at a time, temperature, the context window, why AI makes things up with total confidence, why it is bad at maths, why it agrees with you when you push back, the hidden instructions, and the 2017 paper behind all of it.
When you press send, ChatGPT does not look anything up. There is no library inside it, no list of facts and no saved copy of the internet. Imagine instead a giant web of mushroom roots under a forest floor, where every thread grew thicker each time two words appeared together. Your question lands on it like a spore, and the answer grows from there one step at a time.
CHAPTERS
0:00 Where do the answers come from
0:21 The web of mushroom roots
0:51 How the roots grow
1:32 What it was fed
2:07 The books are gone
2:29 What a thread really is
3:10 Pieces, not words
3:45 Your question is a spore
4:20 It does not know the ending
4:48 How far it wanders
5:16 Same question, different answer
5:46 How much it can see
6:10 A path nobody grew
6:42 Why it sounds so sure
7:12 Why it struggles with maths
7:48 Where the design came from
8:23 Bigger roots
8:56 Teaching it manners
9:34 Why it agrees with you
10:09 The instructions you cannot see
10:45 It is not learning from you
11:15 Asking it to show its working
11:55 The oldest mistake
12:34 The argument that is not settled
13:16 When to trust it
14:00 Back to the send button
SOURCES
Vaswani et al., "Attention Is All You Need", Google, 2017
Brown et al., "Language Models are Few-Shot Learners", OpenAI, 2020
Ouyang et al., "Training language models to follow instructions with human feedback", 2022
Wei et al., "Chain-of-Thought Prompting Elicits Reasoning in Large Language Models", Google, 2022
Bender, Gebru, McMillan-Major and Mitchell, "On the Dangers of Stochastic Parrots", 2021
Sharma et al., "Towards Understanding Sycophancy in Language Models", 2023
Weizenbaum, ELIZA, MIT, 1966, and "Computer Power and Human Reason", 1976
#ChatGPT #AIExplained #ExplainedWithMushrooms #claude