They're so well done. Sadly, he doesn't do them anymore because Youtube's algorithm doesn't make it worth his while. Evidently, he gets the same traffic & revenue from a 10 minute video reviewing "stupid bike gadgets" into the camera as he does for spending a month building a cool bike jump and editing together one of those amazing videos on the playlist.
If youtube rewarded evergreen stuff like that instead of cheap "reaction" videos, it'd be a much cooler place.
He put a video up about this a couple of weeks ago, and rather than YouTube's algorithm being to blame, it was more him struggling to keep up with his audiences constant demand for more progress with more super sketchy features that he was basicly building constantly.
"the Algoirithm" often gets sited, but it's often a youtuber's interaction with thier audience, the pressure to keep up with demand, rising success and not wanting to miss out.
Do they give you the data to know? I assumed he meant that impressions had gone way down so people are not even seeing it (and why everyone begs you to subscribe and hit the bell). Of course impressions could be restricted because people stop watching after 20 seconds.
"The Algorithm" doesn't reward hard work, imho part of the math is our short attention spans... there's only so many time you can watch someone else vlog from a sprinter in Squamish.
Is it the YouTube algorithm's fault? Or viewer preferences? I think YouTube would be happy to recommend the videos to more people if more people watched them.
I watch photography videos on YouTube, and camera review channels consistently have far more subscribers than channels who make content about taking photos. (Or at least they did in the past - in recent years camera tech has really matured and interesting releases are much less frequent, and reviewers seem to have taken a hit).
I think people just like gear. Should YouTube not show people what they like to see?
I've watched some Berm Peak videos in the past and I mostly know the channel for its videos about builds/repairs, or his video about the history of valves. The mountain biking videos are good too, but only hold my interest for so long. If I want to see mountain biking I'm more likely to look at some of the stuff Red Bull is putting out.
I’m gonna say it’s YouTube. They are obsessed with pushing short form shit videos to me despite me never wanting to watch them. I hate YouTube shorts so much
I'm with you on not liking shorts. I read there is a new option to limit time spent in shorts, if you set that to 0 they do go away. Have not tested it.
In the Home Screen it is always completely full of shorts. In my recommendations, full of shorts. There’s no way to just never be shown shorts. It’s very clear from the layout that they REALLY want you to watch shorts. I assume it’s because they are more addictive = more views = more ads. But I pay for YY premium! So why do they still insist on putting shorts everywhere? If I want shitty short videos I’ll use TikTok
But there are ways to never be shown shorts. They just tend to require avoiding the YouTube app.
Remember the World Wide Web? Even here in 2026, YouTube is still just a website -- if you want it to be. Web browsers like Firefox can still be used to block whatever you don't want to see.
So an effective way to deal with this situation is to switch to a browser, block the shorts, and then cancel premium and block ads as well.
In this way: A person can still consume the content that they want, while also protesting with their wallet.
Subscriptions and even viewing history(?!) pull shorts out of order and place them at the top.
Hell, many times I launch YouTube and it immediately starts playing a short, no interaction required. And I have "show touches" on, so I know it's not a phantom tap or something, it's doing it all on its own.
I wonder if they are doing some A/B testing on this. I see shorts about once a month or so, which is when the "not interested" option runs its course. Other than that, they don't push it to me at all, and the app has certainly never started auto playing a short when I've opened it.
I've also never watched a short, or if I did I immediately removed it from my watch history. I do keep that trimmed with prejudice, because the algorithm is desperate to show me something other than what I normally watch.
I somewhat suspect the auto-play is a bug with "accessibility -> remove animations". There are quite a lot of Google-product bugs with it enabled, though most don't render things unusable (except the recent apps list, WOW wtf). Youtube has also lately been jumping straight to "Shorts [that I have uploaded, which is zero]" when I open the "You" section, despite it being the last in the list and completely empty - this too is very likely an animation-related bug.
Why it happens with remove-animations: well I'm pretty confident they don't test with it. But it's super freakin' weird. I kinda doubt it'll get fixed though, doing so will cause "engagement" to drop, and it has been happening for most of a year now.
I don't get them on desktop either, but on the mobile app they're very hard to avoid. Seems like YouTube doesn't want to be YouTube anymore, it wants to morph into TikTok. We don't need more TikToks, having one TikTok is already bad enough thankyouverymuch!
The thing is, "what people enjoy while watching", "what they derive lasting benefit, memory or happiness from", and "what they click on in a thumbnail" are three different things, and youtube optimizes for the latter. Which is why youtube face is a thing.
Click-through rates are indeed very important, but that's not all they optimize for. They are also looking at watch time, what you do after watching the video (do you watch more videos from the same creator, or on the same subject, something totally different, or do you leave the site), whether you interact with the video by liking or commenting, channels you have subscribed to, things you have searched for, etc.
And I think that when you spend a significant amount of time watching videos on a certain subject or from a certain channel - or when you repeatedly decline to watch videos of a certain type when they are suggested - you are signaling a very clear preference.
Are the videos the algorithm serves up something that people will "derive lasting benefit, memory or happiness from"? Probably not - but I also don't think that's what most people are looking for from YouTube. Sometimes, maybe, but more often they are just looking for mindless entertainment. Engaging with media on a deeper level requires effort. YouTube is where people go when even finding something to watch on Netflix is too much effort, let alone doing something healthy.
To keep this all in context - the parent comment was complaining that the algorithm doesn't promote videos of a guy building bike jumps in his back yard enough. I like Berm Peak - but is that something that most people would "derive lasting benefit, memory or happiness from"? No, it's not.
Anyone who hasn't seen those videos hasn't lost out on very much. And for anyone who has spent any amount of time watching videos about bicycles on YouTube - they probably have been recommended Berm Peak videos on numerous occasions. I know I have. The guy has 2.77 million subscribers and 787 million total views on his videos. Whether or not people actually watched the videos when suggested is more likely a matter of personal choice than the algorithm doing him dirty.
On the creator side, YouTube won’t monetize you until you have 1000 subs.
Which means the person that put up 1 video with a massive amount of views will get squat (but Google will gladly put in ads and take a 100% cut).
Protip: subscribe to creators that post useful but not very subscribeable videos. Sucks for creators that put up videos that don’t really relate to eachother.
Dunno why everything has to be a “channel”. That’s what search is supposed to be for.
They must do this as a global level, not an individual level.
I've never noticed a single one of those awful algo-driven apps notice that I never watch short videos and in fact I often force quit the app the second they feed me trash I don't want to see.
There’s a bunch of websites where I’ve declined to install their app about 74747627282 times in a row and… tomorrow they’re still going to think they can nudge me into it.
Yes, modulo only that "what people like enough to keep going through the adverts" isn't exactly the same as "what they click on in a thumbnail", and the latter combines with "did these ads convince someone to get the paid ad free experience" is what YT optimises for.
>* Is it the YouTube algorithm's fault? Or viewer preferences? I think YouTube would be happy to recommend the videos to more people if more people watched them.*
I always got garbage even if it suggested things I wanted to watch too.
The best way is to disable suggestions completely and then just make a note of your favourite channels. That way you get a completely blank landing page and are forced to search out exactly what you want every time.
You can disable suggestions and still subscribe to channels, I'd argue that's the easiest path while still keeping what you're actually interested in. If you have 200+ subscribed channels over the years it's less good, I'd just unscribe if you're not interested in it.
Yup and he talks about how it was unsustainable, and how his priorities have shifted now that he has kids. He does mention the algorithm and the introduction of shorts as something he had to compete with - but mostly he just frames the bike park as an idea that has run its course and he has grown out of.
YouTube mainly rewards people who are doing unsustainable things. It chews them up and spits them out, and when creators either go back to doing something sustainable or crash and burn, the algorithm just stops sending them traffic.
A huge percentage of people who say they are into a particular hobby are really just collectors of that hobby’s gear. Photography is an easy example, but this applies to a LOT of hobbies.
I have long ago realized that I cannot buy more gear for my hobbies unless I commit to using it. I want a lot of cool gadgets, but using what I have needs more time than I put into the hobby. (I need to play mandolin several hours a day to get better, in reality I often skip days or only put in 15 minutes). It is one thing to say "I need X tool for the next step and buy the tool, but I only buy that tool if I really do use it, not because it is a gadget that looks cool (and then I have the tool). I've also found great fun in asking "how did they do this before modern tools" - often I can find an alternate path without the gear.
I remember seeing a video where someone was talking about how from someone on the YouTube team said that their goal is for you to to replace the word “algorithm” with “audience”.
Whether they achieve that goal or not is a different story.
As a young film student my Prof asked me whether I wanted to go to a talk to which he was invited. It was on the genres shown in German public television.
There the summary of the discussion was: Our core demographic are 60 to 70 year olds which is why we only make shows that appeal to 60-70 year olds and our core audience watches TV while doing household chores, so it needs to be simple to follow, so they can do household chores while watching.
I told them that to me this sounds a lot like circular logic, where they justify the things they are doing with the outcomes that produced. It is obvious there are other markets targeting different audiences (e.g. the likes of Netflix have been explicitly mentioned) and these markets are growing based on the way demographics shift.
A bit like a drug dealer that says he can't do honest work since all his customers are drug addicts, they are using the status quo as an excuse to persist the status quo.
The real way to think about these things is to consider them feedback loops. If all your content targets a specific demographic of course you're gonna have more audience members of that demographic, which again leads you to make more content for said demographic, which leads to more audience members of that demographic which... Until you hit some systemic limit, e.g. you have saturated the market or it turns out your content isn't that appealing to begin with in comparison to other stuff.
That means if you want to be strategic about this you need to give incentives to creators to produce stuff for audiences you don't already have. Even better: you need to become a partner these creators can and want to trust in.
These are the levers YouTube needs to pull if they want to stay a relevant platform that people enjoy spending their time on.
I think it's the algorithm. Occasionally I get recommended videos that are 5-8+ years old (so old in terms of Youtube years) with no new comments so presumably not getting a lot of recent views. But soon comes a wave of fresh comments wondering why they never discovered this video before. So the algorithm starts the cycle, not the organic user preferences.
For this particular channel, I watched a bunch of his videos on this Reevo bike In January 2025, and a lot of bike/cycling related videos in general. Despite this clear preference to guide the algorithm, Youtube stopped recommending this channel to me. It disappeared from my feed.
I always suspected Youtube "motivates" creators to pay for promotion by giving them a taste for free, how it looks like to be on everyone's feed, and then takes them off.
Old videos suddenly being heavily promoted again on YouTube is interesting. Sometimes there are obvious reasons, like this 6 year old video about the Straight of Hormuz that is recently making the rounds again [1]. Other times I think it's just the chaotic effects and constructive interference of 2.5 billion users interacting with 15 billion videos - like a rogue wave forming in the ocean.
There are two ways that I've noticed though that YouTube tends to consistently suggest older videos. One is when you first discover a new channel that you like and watch a few videos from that channel, YouTube will start recommending older videos from that channel until you've exhausted the back catalog (or lost interest).
The other way I've noticed is that when I hit the like button a video, YouTube will recommend it to me again after some time has passed. This also seems to depend on the type of video. Music videos are almost always recommended again after some time, while news videos almost never are.
I think these mechanisms are effective at driving traffic to older videos. If I look at my home page right now, most of it does tend to be newer content, but I'd estimate about 25% of it is more than a year old.
In response to your complaint about Berm Peak videos disappearing from your feed - obviously I don't know for certain, but is it possible that YouTube did suggest other Berm Peak videos after you watched a bunch of Reevo videos - but you didn't watch them? And that YouTube might have interpreted that as a lack of interest in the channel?
I've got a couple of channels that I consistently watch whenever they put out a new video. And I find that YouTube is really good about putting their videos at the top of my home page whenever a new video comes out.
> Is it the YouTube algorithm's fault? Or viewer preferences?
I didn’t read the rest of your comment but it’s the fault of the algorithm because that tail wags the dog. It’s physically impossible to watch all videos so we are all at the mercy of the (a) algorithm.
Youtube seems to regularly suggest old videos to me so I think it's less a problem with evergreen content and more that youtube pays for minutes watched so someone who does cheap reaction content can produce more minutes to watch than someone who spends a long time on one video.
I’ve been a fan of his for years (have stickers and a hoodie for berm peak). In the past couple years I stopped watching his channel because he stopped trail building. His new stuff is for entry level people and I’m sure it gets tons of traffic, but watching him make use of his old back yard and his new one was inspiring and fun.
I only saw his previous video on the Reevo that he now fixed (https://www.youtube.com/watch?v=YPe3cY3oeu4 - just in case YouTube doesn't list it under suggested videos when you watch this one) and it was hilarious.
I don't have that model but I can very confidently say :
- do NOT buy an e-bike with custom parts, no matter how "cool" it is!
I bought a CowBoy years ago and honestly, it was great. Until it wasn't. Inexorably, it does not matter how good the parts are, how well designed it is, wear and tear WILL create problems. You WILL need to replace parts. If nobody but the company making the bike sell them you will get in trouble. It's the same with the App, if it's not open source relying on standard AND with existing, not upcoming, GadgetBrige support they will stop supporting your bike and brick it.
Please, pretty please, as consumers who expect to keep on maintaining your bike over years, ASK your repair shop what THEY think is a good bike to fix. Not what is a good bike to ride.
PS: I now ride a Fixie because screw CowBoy and all those e-bike startups who believe they are the new Apple. They aren't and I was the one paying for their delusion.
I have been riding a Panasonic "Gyutto" [1] for about 10 years. I bought it when I had young kids, as it's designed to hold two child seats (one on the back, and one on the handle bars). Now that they've grown up I've replaced the child seats with baskets and use it for commuting every day and grocery shopping.
The thing is an absolute tank - the only parts I've had replace are the tires and brake pads. And the design is really simple with all of the consumable parts being easy to replace. At about $1,400 USD, it's not cheap, but I'm shocked at how long it's lasted and how little maintanence it's needed.
Definitely not "cool" - but one of the best purchases I've ever made.
I think it's a generally good place to start when buying anything, especially anything of high value. You want to be able to truly own it, which includes maintaining and repairing it.
Now of course there are areas you can make trade offs. A lot of people like MacBooks despite them not supporting other operating systems very well and Apple still mostly being hardasses about outside repair, but they come with good performance and battery life.
Making sure to keep maintainability in mind when making a product decision is critical to making an informed purchase.
app is a big point. i was apalled by the fact that the app for the bike (in the video) could no longer connect / authenticate with the bike
i have a e-scooter and judging by the decompiled code it's some sort of chinese e-scooter reskinned by a european company. I know the app is not going to be around for long, so I slowly been trying to make my own.
All the fixie bros say this but my road bike takes basically no maintenance at all. Even when it does, adjusting the derailleurs is a 5 second task with a small Phillips head. Relubing the chain takes 2 minutes. Brakes only needed adjusting when the brake lines were brand new and still stretching and this was also maybe a 30 second task. Rest of maintenance you’d have to do like bb, repacking hub or steerer grease, etc, you’d have to do with a fixie too.
There was a sweet spot where there was. My road bike frame is from the 80s or 90s. It was built for downtube shifting, but now it has brifters. It has mixed shimano group parts from a maybe 2 decade span of time. Quill adapter to use modern handlebars made over the last 2 decades. Pretty standardized 8 speed stuff.
Absolutely insane how people could have bought a bicycle that will become (partially) useless if they don't connect it with a mobile app. Even if this bike have the build quality of a spaceship I wouldn't even touch it with a stick.
Kudos to Seth for cracking the control on the bike, just so we can reclaim control of an appliance that we paid for with our own money, one that won't work because the maker can't be arsed enough to make it work without a mobile app.
A bit of a tangent, but I hate the way this channel uses AI. There's moments where the creator doesn't really know what they're talking about so they ask AI which is a perfectly reasonable thing to do, I don't mind it! However, the way it's repeated verbatim mimicing the speech style of the creator is so offputting I don't really know how to describe it, sounds like someone trying to speak a language they don't know is the best way I can describe it. Might just be me though.
edit:
Actually the comment the creator left on the video is almost purely AI and is just as yuicky to read: "There are lots of questions about runaway Reevo mode, and it's a fun topic. Let's go deeper. First of all, I did add a "PAS Timeout" that turns off PAS after a few minutes, and is selectable in the menu. If you set this up from my Github repo, that feature is active. A lot of you also suggested a weight sensor on the seat, but that would disable pedal assist if you stood up out of the saddle, which is when you would need it most. Another suggestion was the dead man switch. That one would work! This is all fun to think about, so keep the ideas coming. I just MIGHT get another Reevo for myself."
I've been watching Seth for years, that comment is pretty in line with his writing style. I expect he's using AI, but I don't think he would be lazy with it. I would expect him to prefer not replying at all over using AI to reply.
As a very very very long time viewer I've only noticed this in last 10 videos where there's verbatim reads like this from time to time. Ironically where you can notice this the most is the same kind of video: https://www.youtube.com/watch?v=YPe3cY3oeu4
I've seen a pretty big percentage of his videos, dropped off past few years or so, but the writing style is heavily influenced by AI in some parts of the script. It might sound like Seth, but AI cannot replicate it perfectly and it just sounds very off.
It seems to be Claude that was doing the majority of the work on the software, so I find it appropriate that Claude responses (as in Claude-in-a-harness-with-the-task-context-available responses) were passed through. Attribution should be clear though.
someone in the comments theorized it was at first supposed to be a full-fledged front light but the design changed and they just used whatever hardware was already there to power the logo
In the video you can see that the LED lights the logo less evenly than the EL lighting. While that's not important at all, this might be the reason behind this odd choice
I hate how he’s doing the kind of project a ton of people would like to help with, for free and on their own time, yet he’s making an LLM do all the interesting work. He can’t code so he vibecodes the whole frontend. He can’t reverse engineer so he leaves Claude with the uart connection. He doesn’t understand something so he makes an LLM explain it.
With his platform he could easily find a person (or a couple) who could do this work themselves, not only saving him money, but nerdsniping someone into hacking a bike. A win-win for everybody.
Throwing shade at somebody for doing a project with the tools they had at hand is so weird.
This sort of thing exactly aligned with the promise of AI (and every "automation advancement" since the dawn of time)! It's another layer of abstraction that allows less technical people to do the thing.
Imagine after mowing your own lawn with a petrol-powered lawnmower, you got some shades from the neighbors for "Robbing the kids in the community a chance for some honest labor. If you don't know how to use the gardening shear, ask the kids to do it instead of using the automatic lawnmower. Would be a win-win for everybody".
Except with an automatic lawnmower you’re still doing the work yourself. In the video he literally said he left Claude overnight. He did no reverse engineering and no coding
Imagine using a high level language with a mountain of libraries to stitch code blocks together with stack overflow glue, rather than write the project in assembly, fully understand what is going on, and get a much lighter and robust application like a True Programmer™.
You’re either making a bad-faith argument on purpose, or on accident. What you’re saying is false equivalence. Leaving Claude overnight is in no universe comparable to getting some code running yourself, even if that code is 5 lines of python, with a bunch of libraries doing all the heavy lifting
This is what he said though. I don’t doubt he could code if he put his mind to it, but in the video he shows the code that’s talking to the bike, displaying stuff, etc, and says it’s all AI generated because it’s outside his area of expertise.
If you're not familiar with this channel, do yourself a favor and watch your way through his backyard trail build videos:
https://www.youtube.com/watch?v=zAsuk8OndHs&list=PL5S7V5NhM8...
They're so well done. Sadly, he doesn't do them anymore because Youtube's algorithm doesn't make it worth his while. Evidently, he gets the same traffic & revenue from a 10 minute video reviewing "stupid bike gadgets" into the camera as he does for spending a month building a cool bike jump and editing together one of those amazing videos on the playlist.
If youtube rewarded evergreen stuff like that instead of cheap "reaction" videos, it'd be a much cooler place.
In case you want more trail building from Berm Peak:
https://www.youtube.com/playlist?list=PL5S7V5NhM8JSm7EK8IYTS...
https://www.youtube.com/playlist?list=PL5S7V5NhM8JRuAMsVEQ_3...
He put a video up about this a couple of weeks ago, and rather than YouTube's algorithm being to blame, it was more him struggling to keep up with his audiences constant demand for more progress with more super sketchy features that he was basicly building constantly.
"the Algoirithm" often gets sited, but it's often a youtuber's interaction with thier audience, the pressure to keep up with demand, rising success and not wanting to miss out.
Do they give you the data to know? I assumed he meant that impressions had gone way down so people are not even seeing it (and why everyone begs you to subscribe and hit the bell). Of course impressions could be restricted because people stop watching after 20 seconds.
Another once great channel (the loam ranger) did a deep dive into the phenomenon of why mountain bike youtubers were struggling with the old format.
Its nuanced, part safety, partly that its a tonne of work to produce good content when you can simply switch to fluff pieces like Berm Peak
https://www.youtube.com/watch?v=nNiCtWoGonw
"The Algorithm" doesn't reward hard work, imho part of the math is our short attention spans... there's only so many time you can watch someone else vlog from a sprinter in Squamish.
Is it the YouTube algorithm's fault? Or viewer preferences? I think YouTube would be happy to recommend the videos to more people if more people watched them.
I watch photography videos on YouTube, and camera review channels consistently have far more subscribers than channels who make content about taking photos. (Or at least they did in the past - in recent years camera tech has really matured and interesting releases are much less frequent, and reviewers seem to have taken a hit).
I think people just like gear. Should YouTube not show people what they like to see?
I've watched some Berm Peak videos in the past and I mostly know the channel for its videos about builds/repairs, or his video about the history of valves. The mountain biking videos are good too, but only hold my interest for so long. If I want to see mountain biking I'm more likely to look at some of the stuff Red Bull is putting out.
I’m gonna say it’s YouTube. They are obsessed with pushing short form shit videos to me despite me never wanting to watch them. I hate YouTube shorts so much
I'm with you on not liking shorts. I read there is a new option to limit time spent in shorts, if you set that to 0 they do go away. Have not tested it.
In what context do you get Short Form videos? I've never once been pushed to them.
In the Home Screen it is always completely full of shorts. In my recommendations, full of shorts. There’s no way to just never be shown shorts. It’s very clear from the layout that they REALLY want you to watch shorts. I assume it’s because they are more addictive = more views = more ads. But I pay for YY premium! So why do they still insist on putting shorts everywhere? If I want shitty short videos I’ll use TikTok
Yeah, they're very pushy with shorts.
But there are ways to never be shown shorts. They just tend to require avoiding the YouTube app.
Remember the World Wide Web? Even here in 2026, YouTube is still just a website -- if you want it to be. Web browsers like Firefox can still be used to block whatever you don't want to see.
So an effective way to deal with this situation is to switch to a browser, block the shorts, and then cancel premium and block ads as well.
In this way: A person can still consume the content that they want, while also protesting with their wallet.
Subscriptions and even viewing history(?!) pull shorts out of order and place them at the top.
Hell, many times I launch YouTube and it immediately starts playing a short, no interaction required. And I have "show touches" on, so I know it's not a phantom tap or something, it's doing it all on its own.
YouTube pushes shorts insanely hard.
I wonder if they are doing some A/B testing on this. I see shorts about once a month or so, which is when the "not interested" option runs its course. Other than that, they don't push it to me at all, and the app has certainly never started auto playing a short when I've opened it.
I've also never watched a short, or if I did I immediately removed it from my watch history. I do keep that trimmed with prejudice, because the algorithm is desperate to show me something other than what I normally watch.
I somewhat suspect the auto-play is a bug with "accessibility -> remove animations". There are quite a lot of Google-product bugs with it enabled, though most don't render things unusable (except the recent apps list, WOW wtf). Youtube has also lately been jumping straight to "Shorts [that I have uploaded, which is zero]" when I open the "You" section, despite it being the last in the list and completely empty - this too is very likely an animation-related bug.
Why it happens with remove-animations: well I'm pretty confident they don't test with it. But it's super freakin' weird. I kinda doubt it'll get fixed though, doing so will cause "engagement" to drop, and it has been happening for most of a year now.
I don't get them on desktop either, but on the mobile app they're very hard to avoid. Seems like YouTube doesn't want to be YouTube anymore, it wants to morph into TikTok. We don't need more TikToks, having one TikTok is already bad enough thankyouverymuch!
In any context I can think of, even looking at channels I subscribe to on the AppleTV app, shorts are the second row.
I also hate YouTube Shorts. Just put them out on another website.
Luckily I use uBlock Origin and ReVanced, and I blocked all Shorts from even appearing.
> show people what they like to see?
The thing is, "what people enjoy while watching", "what they derive lasting benefit, memory or happiness from", and "what they click on in a thumbnail" are three different things, and youtube optimizes for the latter. Which is why youtube face is a thing.
Click-through rates are indeed very important, but that's not all they optimize for. They are also looking at watch time, what you do after watching the video (do you watch more videos from the same creator, or on the same subject, something totally different, or do you leave the site), whether you interact with the video by liking or commenting, channels you have subscribed to, things you have searched for, etc.
And I think that when you spend a significant amount of time watching videos on a certain subject or from a certain channel - or when you repeatedly decline to watch videos of a certain type when they are suggested - you are signaling a very clear preference.
Are the videos the algorithm serves up something that people will "derive lasting benefit, memory or happiness from"? Probably not - but I also don't think that's what most people are looking for from YouTube. Sometimes, maybe, but more often they are just looking for mindless entertainment. Engaging with media on a deeper level requires effort. YouTube is where people go when even finding something to watch on Netflix is too much effort, let alone doing something healthy.
To keep this all in context - the parent comment was complaining that the algorithm doesn't promote videos of a guy building bike jumps in his back yard enough. I like Berm Peak - but is that something that most people would "derive lasting benefit, memory or happiness from"? No, it's not.
Anyone who hasn't seen those videos hasn't lost out on very much. And for anyone who has spent any amount of time watching videos about bicycles on YouTube - they probably have been recommended Berm Peak videos on numerous occasions. I know I have. The guy has 2.77 million subscribers and 787 million total views on his videos. Whether or not people actually watched the videos when suggested is more likely a matter of personal choice than the algorithm doing him dirty.
On the creator side, YouTube won’t monetize you until you have 1000 subs.
Which means the person that put up 1 video with a massive amount of views will get squat (but Google will gladly put in ads and take a 100% cut).
Protip: subscribe to creators that post useful but not very subscribeable videos. Sucks for creators that put up videos that don’t really relate to eachother.
Dunno why everything has to be a “channel”. That’s what search is supposed to be for.
> but that's not all they optimize for.
They must do this as a global level, not an individual level.
I've never noticed a single one of those awful algo-driven apps notice that I never watch short videos and in fact I often force quit the app the second they feed me trash I don't want to see.
And they keep doing it.
There’s a bunch of websites where I’ve declined to install their app about 74747627282 times in a row and… tomorrow they’re still going to think they can nudge me into it.
Yes, modulo only that "what people like enough to keep going through the adverts" isn't exactly the same as "what they click on in a thumbnail", and the latter combines with "did these ads convince someone to get the paid ad free experience" is what YT optimises for.
>* Is it the YouTube algorithm's fault? Or viewer preferences? I think YouTube would be happy to recommend the videos to more people if more people watched them.*
I always got garbage even if it suggested things I wanted to watch too.
The best way is to disable suggestions completely and then just make a note of your favourite channels. That way you get a completely blank landing page and are forced to search out exactly what you want every time.
You can disable suggestions and still subscribe to channels, I'd argue that's the easiest path while still keeping what you're actually interested in. If you have 200+ subscribed channels over the years it's less good, I'd just unscribe if you're not interested in it.
He actually just did a video about that. https://m.youtube.com/watch?v=E_PdRRfoBXo&pp=ygUJQmVybSBwZWF... It includes a recap of the last several years of content and how he was trying to keep his audience, sometimes by doing really risky things.
Yup and he talks about how it was unsustainable, and how his priorities have shifted now that he has kids. He does mention the algorithm and the introduction of shorts as something he had to compete with - but mostly he just frames the bike park as an idea that has run its course and he has grown out of.
YouTube mainly rewards people who are doing unsustainable things. It chews them up and spits them out, and when creators either go back to doing something sustainable or crash and burn, the algorithm just stops sending them traffic.
A huge percentage of people who say they are into a particular hobby are really just collectors of that hobby’s gear. Photography is an easy example, but this applies to a LOT of hobbies.
I have long ago realized that I cannot buy more gear for my hobbies unless I commit to using it. I want a lot of cool gadgets, but using what I have needs more time than I put into the hobby. (I need to play mandolin several hours a day to get better, in reality I often skip days or only put in 15 minutes). It is one thing to say "I need X tool for the next step and buy the tool, but I only buy that tool if I really do use it, not because it is a gadget that looks cool (and then I have the tool). I've also found great fun in asking "how did they do this before modern tools" - often I can find an alternate path without the gear.
Coincidentally he just made a video elaborating on this as well as a general overview of the history and state of mountain biking channels on YouTube:
https://youtu.be/_oQ4XY4P1n8
I remember seeing a video where someone was talking about how from someone on the YouTube team said that their goal is for you to to replace the word “algorithm” with “audience”.
Whether they achieve that goal or not is a different story.
As a young film student my Prof asked me whether I wanted to go to a talk to which he was invited. It was on the genres shown in German public television.
There the summary of the discussion was: Our core demographic are 60 to 70 year olds which is why we only make shows that appeal to 60-70 year olds and our core audience watches TV while doing household chores, so it needs to be simple to follow, so they can do household chores while watching.
I told them that to me this sounds a lot like circular logic, where they justify the things they are doing with the outcomes that produced. It is obvious there are other markets targeting different audiences (e.g. the likes of Netflix have been explicitly mentioned) and these markets are growing based on the way demographics shift.
A bit like a drug dealer that says he can't do honest work since all his customers are drug addicts, they are using the status quo as an excuse to persist the status quo.
The real way to think about these things is to consider them feedback loops. If all your content targets a specific demographic of course you're gonna have more audience members of that demographic, which again leads you to make more content for said demographic, which leads to more audience members of that demographic which... Until you hit some systemic limit, e.g. you have saturated the market or it turns out your content isn't that appealing to begin with in comparison to other stuff.
That means if you want to be strategic about this you need to give incentives to creators to produce stuff for audiences you don't already have. Even better: you need to become a partner these creators can and want to trust in.
These are the levers YouTube needs to pull if they want to stay a relevant platform that people enjoy spending their time on.
Do you have some recommendations for photography channels on YouTube?
I like Thomas Heaton for landscape photography. For gear reviews I like PetaPixel. And for tips/tutorials I think Simon d'Entremont is good.
There are a bunch of other channels out there too that I watch from time to time, but I think the above are the best in their respective categories.
I think it's the algorithm. Occasionally I get recommended videos that are 5-8+ years old (so old in terms of Youtube years) with no new comments so presumably not getting a lot of recent views. But soon comes a wave of fresh comments wondering why they never discovered this video before. So the algorithm starts the cycle, not the organic user preferences.
For this particular channel, I watched a bunch of his videos on this Reevo bike In January 2025, and a lot of bike/cycling related videos in general. Despite this clear preference to guide the algorithm, Youtube stopped recommending this channel to me. It disappeared from my feed.
I always suspected Youtube "motivates" creators to pay for promotion by giving them a taste for free, how it looks like to be on everyone's feed, and then takes them off.
Old videos suddenly being heavily promoted again on YouTube is interesting. Sometimes there are obvious reasons, like this 6 year old video about the Straight of Hormuz that is recently making the rounds again [1]. Other times I think it's just the chaotic effects and constructive interference of 2.5 billion users interacting with 15 billion videos - like a rogue wave forming in the ocean.
There are two ways that I've noticed though that YouTube tends to consistently suggest older videos. One is when you first discover a new channel that you like and watch a few videos from that channel, YouTube will start recommending older videos from that channel until you've exhausted the back catalog (or lost interest).
The other way I've noticed is that when I hit the like button a video, YouTube will recommend it to me again after some time has passed. This also seems to depend on the type of video. Music videos are almost always recommended again after some time, while news videos almost never are.
I think these mechanisms are effective at driving traffic to older videos. If I look at my home page right now, most of it does tend to be newer content, but I'd estimate about 25% of it is more than a year old.
In response to your complaint about Berm Peak videos disappearing from your feed - obviously I don't know for certain, but is it possible that YouTube did suggest other Berm Peak videos after you watched a bunch of Reevo videos - but you didn't watch them? And that YouTube might have interpreted that as a lack of interest in the channel?
I've got a couple of channels that I consistently watch whenever they put out a new video. And I find that YouTube is really good about putting their videos at the top of my home page whenever a new video comes out.
[1] https://www.youtube.com/watch?v=udtVdDmSSoo
> Is it the YouTube algorithm's fault? Or viewer preferences?
I didn’t read the rest of your comment but it’s the fault of the algorithm because that tail wags the dog. It’s physically impossible to watch all videos so we are all at the mercy of the (a) algorithm.
Youtube seems to regularly suggest old videos to me so I think it's less a problem with evergreen content and more that youtube pays for minutes watched so someone who does cheap reaction content can produce more minutes to watch than someone who spends a long time on one video.
I’ve been a fan of his for years (have stickers and a hoodie for berm peak). In the past couple years I stopped watching his channel because he stopped trail building. His new stuff is for entry level people and I’m sure it gets tons of traffic, but watching him make use of his old back yard and his new one was inspiring and fun.
I only saw his previous video on the Reevo that he now fixed (https://www.youtube.com/watch?v=YPe3cY3oeu4 - just in case YouTube doesn't list it under suggested videos when you watch this one) and it was hilarious.
Or pay for a nebula subscription and watch there.
I don't have that model but I can very confidently say :
- do NOT buy an e-bike with custom parts, no matter how "cool" it is!
I bought a CowBoy years ago and honestly, it was great. Until it wasn't. Inexorably, it does not matter how good the parts are, how well designed it is, wear and tear WILL create problems. You WILL need to replace parts. If nobody but the company making the bike sell them you will get in trouble. It's the same with the App, if it's not open source relying on standard AND with existing, not upcoming, GadgetBrige support they will stop supporting your bike and brick it.
Please, pretty please, as consumers who expect to keep on maintaining your bike over years, ASK your repair shop what THEY think is a good bike to fix. Not what is a good bike to ride.
PS: I now ride a Fixie because screw CowBoy and all those e-bike startups who believe they are the new Apple. They aren't and I was the one paying for their delusion.
I have been riding a Panasonic "Gyutto" [1] for about 10 years. I bought it when I had young kids, as it's designed to hold two child seats (one on the back, and one on the handle bars). Now that they've grown up I've replaced the child seats with baskets and use it for commuting every day and grocery shopping.
The thing is an absolute tank - the only parts I've had replace are the tires and brake pads. And the design is really simple with all of the consumable parts being easy to replace. At about $1,400 USD, it's not cheap, but I'm shocked at how long it's lasted and how little maintanence it's needed.
Definitely not "cool" - but one of the best purchases I've ever made.
1. https://cycle.panasonic.com/products/gyutto_croomr_ex/
Pretty sure at $1,400 it's cheaper than most worthwhile eBikes on the market.
I think it's a generally good place to start when buying anything, especially anything of high value. You want to be able to truly own it, which includes maintaining and repairing it.
Now of course there are areas you can make trade offs. A lot of people like MacBooks despite them not supporting other operating systems very well and Apple still mostly being hardasses about outside repair, but they come with good performance and battery life.
Making sure to keep maintainability in mind when making a product decision is critical to making an informed purchase.
app is a big point. i was apalled by the fact that the app for the bike (in the video) could no longer connect / authenticate with the bike
i have a e-scooter and judging by the decompiled code it's some sort of chinese e-scooter reskinned by a european company. I know the app is not going to be around for long, so I slowly been trying to make my own.
>PS: I now ride a Fixie
Fixies are the way to go. Less parts, less things to break. Very Unix-y.
All the fixie bros say this but my road bike takes basically no maintenance at all. Even when it does, adjusting the derailleurs is a 5 second task with a small Phillips head. Relubing the chain takes 2 minutes. Brakes only needed adjusting when the brake lines were brand new and still stretching and this was also maybe a 30 second task. Rest of maintenance you’d have to do like bb, repacking hub or steerer grease, etc, you’d have to do with a fixie too.
I’ll take my mechanical advantage with gearing.
different spokes for different folks :)
> I now ride a Fixie
With a conversion kit? ^_^
When building my bikes I quickly found out that are no standard parts.
There was a sweet spot where there was. My road bike frame is from the 80s or 90s. It was built for downtube shifting, but now it has brifters. It has mixed shimano group parts from a maybe 2 decade span of time. Quill adapter to use modern handlebars made over the last 2 decades. Pretty standardized 8 speed stuff.
Absolutely insane how people could have bought a bicycle that will become (partially) useless if they don't connect it with a mobile app. Even if this bike have the build quality of a spaceship I wouldn't even touch it with a stick.
Kudos to Seth for cracking the control on the bike, just so we can reclaim control of an appliance that we paid for with our own money, one that won't work because the maker can't be arsed enough to make it work without a mobile app.
Related: Cory Doctorow's [Unauthorized Bread](https://arstechnica.com/gaming/2020/01/unauthorized-bread-a-...)
A bit of a tangent, but I hate the way this channel uses AI. There's moments where the creator doesn't really know what they're talking about so they ask AI which is a perfectly reasonable thing to do, I don't mind it! However, the way it's repeated verbatim mimicing the speech style of the creator is so offputting I don't really know how to describe it, sounds like someone trying to speak a language they don't know is the best way I can describe it. Might just be me though.
edit:
Actually the comment the creator left on the video is almost purely AI and is just as yuicky to read: "There are lots of questions about runaway Reevo mode, and it's a fun topic. Let's go deeper. First of all, I did add a "PAS Timeout" that turns off PAS after a few minutes, and is selectable in the menu. If you set this up from my Github repo, that feature is active. A lot of you also suggested a weight sensor on the seat, but that would disable pedal assist if you stood up out of the saddle, which is when you would need it most. Another suggestion was the dead man switch. That one would work! This is all fun to think about, so keep the ideas coming. I just MIGHT get another Reevo for myself."
I've been watching Seth for years, that comment is pretty in line with his writing style. I expect he's using AI, but I don't think he would be lazy with it. I would expect him to prefer not replying at all over using AI to reply.
If you know Seth you know he kind of talks like that comment reads, and has for quite a while
As a very very very long time viewer I've only noticed this in last 10 videos where there's verbatim reads like this from time to time. Ironically where you can notice this the most is the same kind of video: https://www.youtube.com/watch?v=YPe3cY3oeu4
You don't know much about Seth do you? That's not "AI", it's called a script.
I've seen a pretty big percentage of his videos, dropped off past few years or so, but the writing style is heavily influenced by AI in some parts of the script. It might sound like Seth, but AI cannot replicate it perfectly and it just sounds very off.
It seems to be Claude that was doing the majority of the work on the software, so I find it appropriate that Claude responses (as in Claude-in-a-harness-with-the-task-context-available responses) were passed through. Attribution should be clear though.
This comment doesn't seem obviously AI-written to me.
Maybe the trick is to own it - are AI vtubers already a thing?
Neuro-sama has been a thing for quite a while, and is much more interesting than ChatGPT in a waifu suit. https://youtube.com/channel/UCLHmLrj4pHHg3-iBJn_CqxA
I can’t imagine having such a functional view of “consuming content” in my life that I’d ever want to watch that.
Can anyone explain why in the world they would use an electroluminescent light instead of a simple LED?
someone in the comments theorized it was at first supposed to be a full-fledged front light but the design changed and they just used whatever hardware was already there to power the logo
You don't use EL as a "full fledged front light", they don't have the lumen output and need a lot of support (high voltage) to produce what they do.
In the video you can see that the LED lights the logo less evenly than the EL lighting. While that's not important at all, this might be the reason behind this odd choice
I have a red one and a blue one. In storage at the moment.
I guess I won't be riding them anytime soon. But I am glad to know there is a way to resurrect/improve them!
If anything this video is kicking off a storm of amateur hacking enthusiasm for these wild bikes. Do us all a favor and sell yours!
ok AND
I hate how he’s doing the kind of project a ton of people would like to help with, for free and on their own time, yet he’s making an LLM do all the interesting work. He can’t code so he vibecodes the whole frontend. He can’t reverse engineer so he leaves Claude with the uart connection. He doesn’t understand something so he makes an LLM explain it.
With his platform he could easily find a person (or a couple) who could do this work themselves, not only saving him money, but nerdsniping someone into hacking a bike. A win-win for everybody.
Throwing shade at somebody for doing a project with the tools they had at hand is so weird.
This sort of thing exactly aligned with the promise of AI (and every "automation advancement" since the dawn of time)! It's another layer of abstraction that allows less technical people to do the thing.
Imagine after mowing your own lawn with a petrol-powered lawnmower, you got some shades from the neighbors for "Robbing the kids in the community a chance for some honest labor. If you don't know how to use the gardening shear, ask the kids to do it instead of using the automatic lawnmower. Would be a win-win for everybody".
Except with an automatic lawnmower you’re still doing the work yourself. In the video he literally said he left Claude overnight. He did no reverse engineering and no coding
this is berm peak though, coding was never going to be the focus of the video
It shows that you don't need to know how to code to succeed in this type of project.
As someone who does know how to code, I find the approach to be great, as it can motivate others to try similar projects.
Imagine using a high level language with a mountain of libraries to stitch code blocks together with stack overflow glue, rather than write the project in assembly, fully understand what is going on, and get a much lighter and robust application like a True Programmer™.
You’re either making a bad-faith argument on purpose, or on accident. What you’re saying is false equivalence. Leaving Claude overnight is in no universe comparable to getting some code running yourself, even if that code is 5 lines of python, with a bunch of libraries doing all the heavy lifting
I'm pretty sure if he did that someone would complain that he is using people as free labour to increase his youtube revenues.
It's impossible to do anything on the internet without someone in the peanut gallery telling you you are doing something wrong.
He used to be a web developer before doing youtube fulltime, so "he can’t code" is a false premise.
This is what he said though. I don’t doubt he could code if he put his mind to it, but in the video he shows the code that’s talking to the bike, displaying stuff, etc, and says it’s all AI generated because it’s outside his area of expertise.
Yep, we're becoming further and further apart by the day.