It was impressive how he spotted a bug in his rendering engine and found a workaround for it in realtime!
It's funny how, after years of hearing his voice and not seeing his face, seeing his face puts me in smack in the middle of the uncanny valley.
I feel like quite a few of the content creators I watch are starting to “show their face”, though the big one that comes to mind is the Real Engineering guy. A few of his newest videos have had him acting a bit like a host, interviewer and narrator.
It’s also weird when some of these creators swap with somebody else, like veratasium has had his producer do some of the videos instead.
The big reveal would be AvE.
His commencement addresses might really freak you out, then! See e.g. https://youtu.be/z7GVHB2wiyg
Especially because when narrating the audio is better than someone being filmed casually, that small difference can get really weird
He featured multiple times before with Matt Parker and Brady Haran (NF), strange that you missed it. Although I guess even in this niche channel preferences vary wildly.
I’m amazed at how much production goes into each of his videos. His YouTube play button is well deserved.
It's also what bums me out about YouTube. There is an insane amount of effort that goes into producing high-quality videos - orders of magnitude more than would go into putting together a well-illustrated blog post.
As with blogs, a lot of this effort is wasted unless you get lucky. But with blogs, at least you have multiple shots at visibility. Maybe you'll make it to the top of HN, maybe on X, maybe somewhere else. Even within a single platform, you usually have multiple tries. If you don't get noticed right away, there's still hope that someone else shares your content down the line.
In contrast, on YouTube, an algorithm essentially decides once. If you don't already have a zillion subscribers, it shows your video to a couple of people, often more or less at random. If they don't engage, that's the end of the road.
It's such a weird system. It's capable of producing breakout wonders that wouldn't be possible pre-YouTube, but it's also so punishing, and so much great content goes there to die.
A YouTube video has a URL though. So just like a blog post, you can share it on all the same sites you mentioned with blog posts.
Plus you have the built-in audience of YouTube and the algorithm that can help with discovery..
"Build it and they will come" has never been true, for videos or blogs...
There are surprisingly few venues for video content outside YT, at least not on a scale that would matter on YT! For example, if you want to get to the top of HN, non-video content has much better odds. Many tech- or science-centric subreddits on Reddit discourage or ban videos too.
The point is, YT is a fairly closed ecosystem. And because of how their algorithm works, for every 3Blue1Brown, there are countless people who put days or weeks into great science visualizations and get 50 views.
Omg I would love to make an explainer on bridging algorithms[1] with this tool! I've been a huge fan of their use in participatory democracy processes since 2016 (using tools like Pol.is), and have wanted to make a contribution to increasing literacy around the foundational math involved <3
Had I known about Manim back when his Summer of Math Exposition[2] was happening, I def would have dived in!
What is the actual math involved? From following the link in [1] I found almost no math content.
I'd watch if you made it. Big fan of pol.is : )
This seems like some kind of NGO word salad slop dressing itself up with a bit of extraneous math notation; this is almost exactly the opposite of the type of stuff 3B1B covers in his videos, or that (I must imagine) Manim is optimized for.
I assume that a considerable percentage of CS students or recent graduates (myself included, back in the day) dream about creating some sort of visualization tool, let alone an awesome one like this.
Sometimes knowing how to do something isn't nearly as important as badly wanting to do something
For some it can be far easier to find knowledge than it is to find motivation
Link to Manim: https://github.com/3b1b/manim
and the community fork: https://github.com/ManimCommunity/manim
Wow. I really wish this were a JavaScript library so we could play with this in a browser and publish 3D animations to the browser.
Check out Motion Canvas[0] by Jacob from aarthificial, prolific coder and animator [1].
I've done some porting between Python and JS based on Tensorflow in the past - and I suspect the poor ergonomics in JS for math/lists would probably ruin the experience a good amount.
Perhaps something like Pyodide can bridge the gap and make it easier to bring into the browser as well.
It would probably be easier to add an 'export to js' option to Manim than to try to create an ergomic js library.
Motion Canvas might be a great fit: https://motioncanvas.io/