In this vlog, Andrej Karpathy celebrates reaching 4,096 YouTube subscribers and thanks his audience. He explains why recent uploads have slowed—vacation, reduced cubing practice, and a desire to keep his channel as the go‑to source for 3×3 speedcubing. He critiques Pogobat’s popular beginner tutorial, argues his own Beginner’s Method video is superior, and asks viewers to rate his video to improve its YouTube search ranking. Karpathy announces that his website, badmephisto.com, has been moved to a better hosting service, making it faster and more reliable. Looking ahead, he mentions needing better PLL algorithms for his site, is open to programming tutorial ideas despite feeling inexperienced as a teacher, and is launching a separate channel focused on accessible mathematics such as chaos theory, Julia and Mandelbrot sets. Finally, he demonstrates a collection of “useless” skills he’s been picking up—solving a Sudoku cube, juggling, dice stacking, yo‑yo tricks, pen spinning, and solving the 15‑puzzle—showcasing his hobbyist side before signing off.
Chapters
Hi guys! A couple of weeks ago I reached another subscribers threshold, when I reached 4,096 subscribers. I was not able to make a video on it right away because I was on a vacation. I was also busy gathering other useless skills, which I will show you at the very end of the video. But first, I wanted to share my thoughts on this channel and talk about some other administrative things. First and foremost, I have to thank all of you for the very nice words and comments, which I get in my mail daily. I am naturally very happy that my channel got this popular and that my efforts are appreciated. Now, you may have noticed that I have not been posting too many videos recently, and there are two main reasons for this. First, I must confess that I have not been cubing too much lately, and nor did I try to improve in some way by learning something new and exciting that I could later on report on. And second, I wanted my channel to be the go-to place for all 3x3 frigicubing, and I feel that I have achieved that to a large extent. And even though I am a strong perfectionist, I can say that I am proud of most of my videos. Which brings me to my second topic that I wanted to discuss. I worked pretty hard on my latest Beginner's Method video, and I think that it is pretty good. Maybe a few people would dispute that and argue that some other video is better or just as good, but I think everyone that knows anything about speedcubing will agree that it is at least much better than Pogobat's video. For those of you who are not familiar with Pogobat's video, it is basically a video that received about 11 million views because it was once featured on the main page of YouTube, but despite that, it is a badly structured video that goes way too fast and confuses many people. And you can see this easily, by the way, in the comment section. I'm not just making it up. The beginner's method he teaches is also very bad in itself. In other words, the video just discourages people to get into speedcubing by both making it hard to learn and also making it hard for people to switch to Fridrich because the method is all screwed up. Now, what worries me is that if you currently go to search "how to solve a Rubik's Cube" here on YouTube, PogoBat's video will come up as the number one hit. A video's place in this list is determined by several variables, but among other things, ranking makes a huge difference. Now, those of you that have been following me for a long time know that I have never asked people to rate my videos. So this may sound ridiculous, but I think for the first time ever, I want to ask you guys explicitly to please rate my beginner's method video, if you thought that it was good, as a service to humanity. I will post the link in the video description. I had a lot of people tell me that they wish they had found my video instead of Pogobad's, and I believe them. I wish we could get the YouTube search algorithm to realize this. Next, some good news. Maybe you have noted that my website badmephisto.com was sometimes very slow and sometimes even unavailable. Now, I've been making a few cents a day off the advertisements on the sides of my videos, and I have now used the money to purchase a proper hosting service. Hosting solution. So the website should now be much faster and responsive. And lastly, my plans for the future. Unfortunately, I do not currently have any big plans for future cubing videos, but I do plan to work a little on my website. In particular, I really need to get better PLLs because I think some of my PLLs suck. If you think you know a very good PLL page, please link it to me because I really need to get some new algorithms to put on the website. Many of you have also suggested that I do some programming tutorials. I am very glad that I have inspired some of you to try to pick up programming, but I doubt a little that I can do a good job of explaining it, because I've been doing it for a very, very long time, and I take a lot of things for granted. But I like challenge, so I may end up making a channel for it. Stay tuned. Although, I have been silently making a few videos on mathematics. I'm pretty fascinated by chaos theory and fractals, so I made a new channel where I teach people for people who never did mathematics before about stuff like the Julia sets, the Mandelbrot set, the complex numbers, stuff like that. Many of these things are not only very popular and interesting but can also be easily explained, so I may continue to make more videos on some cool but easy mathematics on that channel, so subscribe if you are interested. Link in the video description. And lastly, as promised, I'll show you some other useless skills that I've been picking up. So Rubik's Cube, you know, yeah, it gets old. No, I still love you. But anyway, I get a Sudoku cube and it's much harder. It's also much harder to turn, but that's not why it's harder. But the 5s, they can face— they can be twisted the wrong direction. It's driving me crazy. And also the numbers, they're not color-coded. I wish they were because it's really hard to see what goes where. Then I've been picking up some juggling And, ah, I don't think I can demonstrate. Anyway, I can juggle these 3 balls, and I know several tricks, but obviously I can't demonstrate here. Then I've been picking up, I don't know if you guys know this, dice stacking. And I just really recently picked this up, so complete fail, obviously. But I will look into that later. I was trying out things already. Then I've been picking up Yoyoing. So I can do a few tricks and stuff like that. It's pretty cool, I like it. I like to play with it. Obviously I can't demonstrate again on this table, but I can do it, trust me. Then I've been picking up pen spinning. So this is the only trick that I know. Yeah. Oops. Okay, so, so yeah, so complete fail still. Well, not really. And I can do it the other way if I try really hard. Oh! Okay, maybe not. But this is a bad pen. And lastly, I have to show this on my iPhone. A 15-puzzle. And I love these things. This is so much fun. So I've been solving this for a while now. So see, you have to arrange all the numbers: 1, 2, 3, 4, 5, 6, 7, 8, up to 15, and then the puzzle is solved. So I can do these in like 20-30 seconds now, which is pretty good. So 7-8, 10-9, 13, oopsie, 14, and now finish up. Ta-da! I'm a winner! Anyway, so that's what I've been up to. Bye-bye. In particular, I really need to get better PLLs because my PLLs suck. They suck. Suck suck suck. This video has to be deleted.
