yego.me
💡 Stop wasting time. Read Youtube instead of watch. Download Chrome Extension

How AI, Like ChatGPT, *Really* Learns


2m read
·Nov 7, 2024

The main video is talking about a genetic breeding model of how to make machines learn. This method is simpler to explain or just show. Here is a machine learning to walk, or play Mario, or jump really high. A genetic code is an older code, but it still checks out, and I personally suspect in the future genetic models will have a resurgence as compute power approaches crazy pants.

However, the current hotness is deep learning and recursive neural networks, and that is where the linear algebra really increases and explainability in a brief video really decreases. But if I had to kind of explain how they work in a footnote, just for the record, it's like this: No infinite warehouse. Just one student. Teacher Bot has the same test, but this time Builder Bot is 'Dial Adjustment Bot,' where each dial is how sensitive one connection in the student bot's head is.

There's a lot of connections in its head, so a lot of dials. A LOT, a lot. Teacher Bot shows Student Bot a photo, and Dial Adjustment Bot adjusts that dial stronger or weaker to get Student Bot closer to the answer. It's a bit like adjusting the dial on a radio. Is that still a thing? Do cars have radios still? I don't know, anyway.

You might not know the exact frequency of the station, but you can tell if you're getting closer or further away. It's like that but with a hundred thousand dials and a lot of math, and that's just for one test question. When Teacher Bot introduces the next photo, Dial Adjustment Bot needs to adjust all the dials so that Student Bot can answer both questions. As the test gets longer, this becomes an insane amount of math and fine-tuning for Dial Adjustment Bot.

But when it's done, there's a student bot who can do a pretty good job at recognizing new photos, though still suffers from some of the problems mentioned in the main video. Anyway, that's the most babies' first introduction to neural networks you will ever hear. If it sounds interesting to you and you like math and code, go dig into the details; machines that learn are the future of everything.

Maybe, quite literally, the future of everything, and given what we've put them through, may the bots have mercy on us all.

More Articles

View All
Water Technology in Architecture | National Geographic
[Music] Here on the snowy slopes of Mount Hood, Oregon, it seems impossible that the U.S. could ever run low on water. But government-backed research says we could in little more than 50 years. [Music] Oregon relies heavily on snowmelt for its fresh water…
New Technologies: Making Wildlife Cinematography More Accessible | National Geographic
[Music] I always wanted to go and explore far away in empty places. From very early on, I just wanted to travel and discover places that weren’t impacted by humans. We have got on 1.6 inside the heart. After several years as an Antarctic ecologist, I had…
Homeroom with Sal & John Stankey - Tuesday, December 1
Hi everyone, Sal Khan here from Khan Academy. Welcome to the Homeroom live stream! We have a very exciting guest today: CEO of AT&T, John Stankey is here. So start putting your questions on Facebook and YouTube, wherever you’re watching it, and I will…
Arrogance & Pride in Stoicism | Q&A #4 | June 2019
Hello everyone, welcome to the QA of June 2019. How are you all doing? Man, oh man, it’s been so hot the last few days in the Netherlands! I understand why they take afternoon naps in some warmer countries, because when it’s above 30 degrees outside, you …
Bank balance sheets and fractional reserve banking | APⓇ Macroeconomics | Khan Academy
In this video, we’re going to talk about balance sheets, and in particular, balance sheets for banks and a fractional reserve lending system. Now, it’s not just banks that have balance sheets; all corporations have a balance sheet. You can even have your …
Example: Comparing distributions | AP Statistics | Khan Academy
What we’re going to do in this video is start to compare distributions. So for example, here we have two distributions that show the various temperatures different cities get during the month of January. This is the distribution for Portland; for example,…