A pattern you'll often find on YT is similar to how Intel and other manufacturing companies do the "Tick" and "Tock" pattern. They'll release a difficult to research, complex and time consuming video, followed by a few easier to produce videos with a lower cost of entry knowledge wise to pay the bills.
Tom7 is an interesting exception to this rule. He puts out about one video per year, but they're all very very well done. But, they're also tongue-in-cheek and he's not _really_ trying to teach you anything - but you'll probably learn something in the process!
It is well understood, and psychologists will likely say the same, that the outcomes from people when they are intrinsically motivated generally surpasses that from extrinsic motivations. We have taken the web away from the former and handed it to the latter and in doing so we deprive ourselves of wonderful things many people have to offer. The web doesn't need paid content creators to survive. There are a huge number of people who are interested in participating for epistemic or community reasons and we need to safeguard these interests. I don't think it can be done in the web as it stands today. I think we really need to go radical and carve out a separate space for such interests. We need a new web.
It's not like Youtube makes you pay for hosted videos if they don't get monetized properly. So in a way those who are properly monetizing Youtube are helping to pay for a free platform, no?
You may argue that such "deep value content" won't be boosted by the algorithms, and I don't see the problem of that either since it's such a niche content that wouldn't even make sense to be considered popular by the algorithm standards. Now you even have the "Bell Notifications" to get notified about new videos, so if that niche content finds true fans why can't they simply just do that?
You have plenty of channels like that, and they still grow.
Maybe the problem is the motivation for those who make such niche content that don't get the proper monetization, but that it's mostly a problem of the audience (not all niches are worth the same, because people have different behaviors and spending habits, and that's fine). That's why the "finances" niche is more valuable then "programmers" niche (or at least was some time ago, now I don't know).
Or am I thinking about this in the wrong way?
When the deep value creators are fighting for attention/views against "view efficiency optimizers", the deep value loses out, making it harder for them to make a living.
Same reason quality journalism is on decline. It's a hard problem to solve as it occurs naturally in a competitive environment.
I understand that, and like you said, it's a hard problem to solve - and it's not a new problem. TV networks experienced the same thing, and that's why part of their models are successfully replicated by Youtubers.
Edutainment is a good example of it, where you could engage the audience into learning while they have fun. So it seems like there's a mid term where education and entertainment feed off each other and pleases a wider audience - and back then the only algorithm was zapping.
Because for the vast majority, probably, deep value is "boring", and that might even be amplified by the motivation behind using Youtube (which I dare to say is mostly for entertainment).
Even for those who find entertainment and extreme value from "deep value creators", they might need to set a bit of time to be engaged to consume that content, and in the end if their "CPM" is not as valuable as other audiences "CPM" (because they engage less with ads, use adblockers, spend less money after engaging with ads, etc), the creator will get less value from those views... and that's just one of those things: it is what it is.
You can't "blame" the audience, or the content creator, or even Youtube.
In the end this revolves around value transactions. Maybe a solution is for the content creator to try to harvest the value from his "deep value created" directly from the audience since Youtube might not be getting that value out of them.
A monetized YouTube buries quality under orders of magnitude more shit than a non-monetized one does. Without the ads, you'd get a thousand good videos among ten thousand crap ones; now you get ten thousand good ones among ninety billion crap ones.
If your video does not retain the viewer, they drop off, this signals to the algorithm not to suggest the video to others.
So for content creators, you spend quite a bit of production time making a long well made video. You get discouraged when you put in quite a bit of effort and it does not do well on the YT platform.
Even for something like Azure Devops I managed to learn much more quickly through YouTube than by reading the docs.
It's a akin to the difference between a book and a painting or a piece of music. A book -- or more generally the written word -- is really good until it isn't. Sometimes knowledge just flows more easily through other media.
Though I do agree that GUI-heavy processes like AWS are better presented in video so you can find exactly what to click on and what happens next.
Cost of production is close to zero. Sit in front of a computer and talk while you work. No edits required, upload as soon as you're done.
- Primer: mainly about simulations and genetic algorithms - https://www.youtube.com/c/PrimerLearning
- Reducible: about computer science and algorithms in general - https://www.youtube.com/c/Reducible
- The Chermo: about game dev in C++ - https://www.youtube.com/c/TheChernoProject
- Sebastian Lague: about game dev/simulations with Unity - https://www.youtube.com/user/Cercopithecan
Channel link: https://www.youtube.com/c/HusseinNasser-software-engineering
IMHO the best strategy to get high-quality content for specific topics is to look for conferences that post their talks. Personally I enjoy e.g. the CppCon and RustCon channels, as well as the PyCon one. There are some other sites like media.ccc.de where you'll find a ton of technical videos as well.
The channel covers a ton of topics but the viewership is low and there's no clear trend in video popularity. Therefore you don't know if a video is any good until you get a few minutes in, but some are very good. And they do deliver on the technical quality front.
YouTube's whole deal is that (1) they will promote you only if lots of people consume your content immediately and rabidly and (2) they will pay you a tiny bit for each view, so you need a lot of views to make anything.
This makes it nearly impossible for an author to make in-depth content on niche topics while making a profit. The audience is (1) too small for YouTube to care and (2) often only needs your niche educational content once, so they don't become rabid, multi-year subscribers. The lack of continued engagement snowballs to make YouTube promote your content less and less, so you never get views. Even worse, every niche video will cause YouTube's algo to actively make each future video on your channel get less and less views - a depressing downward spiral for your channel.
You might say that someone might make in-depth coding content for free (and a few do, like Coding Secrets), but the kind of people who can communicate well and also program well are in huge demand and probably too busy making real money to spend hundreds of hours to make $3 on YouTube when much better options exist.
On the other hand, this is way you see so many YouTube ads for 3rd party paid online course services. These services (LinkedIn Learning, SkillShare, etc.) pay authors more reasonably for their niche content and are built around promoting content for audiences who want to learn. The other strategy that works well is for the author to make very general content on YouTube and then advertise their own course or book to their biggest fans. Both models can sustain someone's full-time income or author's who build an audience and then self-publish their own content can often make 6 figures+.
There are also great meetups that share their video online.
You might also consider some of the free courses on YouTube. I've found Harvard's and Yale's courses to be optimized well for video, MIT's less so (it depends on the course though).
I don't really follow a lot of the higher level abstraction channels like "Computerphile." Although I do generally enjoy those videos when I see them.
I'm more likely to search for the topic that I want to learn about and then skip around many videos until I find the one with the production quality, tone, volume, etc that is most clear and enjoyable to me. Often, you do have to settle for badly made videos on obscure topics (but no complaints! Better than nothing.)
BTW, I'm not disagreeing with you at all as I too agree. I'm just wondering what things you also see as missing.
I find most user content is focused on the introductory level, which makes sense if you are attempting to appeal to the largest audiences for income (so not faulting anyone). I often see content on data engineering but most just define "data engineering" and at most talk about things high level Hadoop, Spark, etc. and do not provide any examples/patterns of working with data that is even close to representative to what I do in my day job. The same is true many other technical disciplines too. With a couple notable exceptions on some security topics (such as demonstrating certain exploits), and some hardware topics -- such as Ben Eater's great work (https://eater.net/)
and of course my channel!:
Lengthy, in-depth technical content doesn't match those criteria.
So if you want to produce that kind of videos, you end-up with only 2 options:
- Channels that are not motivated by revenue, that will post a series of videos then slumber or get abandoned.
- Channels that are sponsored by companies that will try to sell you their language, framework, SaaS, etc... masquerading as technical content
IDK, I think this might be better listed as an "awesome-x" list
Paid is really the way to go. O’Reilly has a lot of great technical content in video form. They also have live learning sessions.
I'm not a fan of videos for learning anything technical. I'd prefer write ups, illustrations and photographs. Videos can be useful sometimes, but what it usually winds up doing for me is slowing down my pace.
A lot of the difficult problems at work are specific to the constraints we have so I would assume the same case for nearly everyone else. Most of these problems have nothing to do with the tech stack or specific language features. In software development the actual programming is the easy part, integrating systems and dealing with scale are the harder parts of the problem.
There are actually some really good YouTubers on some really niche topics. For instance I've found Rahul Nath's videos on Azure Pipelines to be exceptionally helpful: https://www.youtube.com/user/rahulnathp
Have you tried searching for content specific to the niche you're interested in? Chances are if it's a topic that's popular on StackOverflow and that's also visual in nature, there's a good Youtube video explaining it. If it's not visual in nature, the probability of your finding a good video decreases because video is likely not the best medium for conveying that information, which decreases the view potential and the incentive to monetize.
Compare that to the three-part introduction to sound synthesis from the 80s where the dude twiddles knobs for three hours and explains what each of them is doing. Just dialing in one sound takes him ten minutes. It's the best educational video I've seen on YT: https://youtu.be/atvtBE6t48M
The Applied Science channel is a bright counter example, but these are quite rare.
Maybe that's because of how many companies have stopped caring if candidates are actually good software developers and just do leetcode tests and call it a day.
Imo text should be the main source, then use YT videos as supplemental material
Text has its own corresponding set of unique "platform" problems.
But really, the thing spanning both is the concept of "spam". Quantity over quality seems to be winning out, and the algorithms are not doing their job of sorting the best content to the top. They do to an extent, but then they "spread" out the rest over the entire stack of quality so that it's not a winner takes it all kind of sorting.
This channel maybe the best in terms of quality about low level design. If you are a programmer make sure to checkout his low level design playlist: https://www.youtube.com/c/anomaly2104
This is a quality channel to get an intuition for algorithmic thinking: https://www.youtube.com/c/vantonspraul
This is a pretty advanced Rust channel: https://www.youtube.com/c/JonGjengset
If you are a Frontend developer and you want to understand the primitive stuff, how things work from inside do checkout this channel, contains lots of problems and solution including implementation of vDOM, polyfills for Promise, Async Await and various JS stuff:
These two are nice intro channels to get into rust:
If you like maths, do checkout this channel:
This channel also uploads rarely but the quality of videos is very high, it contains general news and stuff related to Science(Physics, Chemistry, Bilog:
If you want to understand modern web layout give this channel a try:
If you are any how involved in Crypto, this is a must to subscribe:
I see I have posted tons of channels but I think these are really good, in terms of quality of content.
Also you can combine the keyword + reddit the results are also good in terms of quality. Often time you find few discussion which provide good learning points.
E.g. one of my faves: https://youtu.be/ajGX7odA87k
But I think you are right, that videos about programming made explicitly for YouTube just aren’t that great.
Numberphile is great. Computerphile is … sometimes good. Which makes me think there is just something about YouTube that doesn’t lend itself well to programming.
For example currently leading a big technical project where all the tech is new to me. Just went on youtube and searched for "$the_tech_I_want_to_learn 2021" and found playlists of quality videos (like an hour long) where the presenter goes very deep into these topics.
Some videos I've found are also by authors of popular books of $the_tech_I_want_to_learn, so that's also awesome to get their insights outside of text.
Just avoid "big channels" because just like anything else at a large scale they compromise on many things. Find the channels that are focused on sharing technical information, not milking the monetization system.