Also, these comments really really really need threading.
Also, these comments really really really need threading.
I may not have as loud a voice as the other people in this conversation, but I too am excited about the future of Cake. For similar reasons as @Scobleizer, Chris as long as I have known you in my partnerships with SmugMug you have been a kind and thoughtful visionary. You have always taken the time to ask the little guy for their input, and taken the time to create a meaningful experience for everyone. Not all of the great ideas you've had got the chance to be realized, but I'm fortunate enough to have seen behind the curtain a bit to know the kind of company and community you want to build.
What brought me to the Internet at a young age were the communities. It was being able to log into a forum in CompuServe on a topic knowing that's where my people were. Every single social media platform today has forced us to engage too much in the bigger, louder world. It's insisted that every person has an equally qualified voice on every subject, and as much as it's important to give people a platform to express themselves, sometimes we need to be a bit more deliberate in how we start a conversation. Panel conversations are exactly what we need right now.
I've been on a bit of a hiatus from things for a while, but I am excited to be back and am ready to invest my time in Cake and see it grow and thrive. I know it will.
Thank you for a great insight about following topics and people, David!
Fundamentally we built Cake to be all about following topics, but as the platform grows there'll be many people talking about the same topic at various levels of credibility. This will create a lot of noise. There are a few ways to solve this problem:
1. Prioritize showing content from people you care about in a specific topic. This solves the problem of following people and seeing everything from them, rather than what you both share interest in.
2. Panel conversations is another solution. These are public conversations that can be read and reacted to by any one, but only panelists can post. This cuts down the noise, prevents trolls and creates a safe space for the panelists to talk.
Here is an example of a recent Panel:
Hope to see more insights from you here on Cake.
I'm curious how Cake will solve THAT particular problem.
Thanks, Chris. That problem is my second-worse nightmare. #1 is no traction at all. 😳
I think all the big services were built by very young men in Silicon Valley around 15 years ago when we were all obsessed by free speech, no? On the upside, we've all been awestruck at the power the networks have to make things go viral.
I heard Ev Williams say if they'd known about the rise of trolls back then, they would have architected Twitter differently. The question we asked was okay, we have the opportunity to learn from 15 years of history. How to make something that will scale?
Once again I thought of Jeff Bezos, who said it's important to ask what will never change. In his case, it's selection, fast delivery, and low prices. They had to obsessively shoot for those without adding bad products, cluttering search so you couldn't find anything, and they had to design warehouses like no one had ever imagined before.
With us it's signal-to-noise and content discovery, no? We have to obsess over those at scale. We've done a few things so far but we're early and the team is small. But focusing on those two things is really energizing.
For example, we don't care about how many users have registered. But we do care about how we get great conversations that get lots of views. In that way we feel more like YouTube than the others, at least to me. One YouTube creator can create millions of views with epic content. Or you can make a great video on how to fix a certain model water heater that a small but devoted group of people love. And they do a really good job of surfacing interesting music and videos, it seems to me.
Trying her out. Like the clean look, so far. One concern I have, though, is reading that Cake is designed primarily for people to follow topics, not people. That's great, in theory, but I'm not sure I understand why both models aren't supported.
Hey David, I think you're absolutely right and we've been thinking this through for awhile now. I think what you want is to follow topics like Tesla Model 3, and when you see someone posting in that topic who is really knowledgeable, you want to follow them in that topic. Am I right? Or possibly in all topics you follow, but not in politics, raising children, or thoughts on the keto diet, which you don't follow. Correct?
It seems to me the problem of social graph companies is you get everything your uncle posts. You follow him to keep up with the cousins on Facebook, but he has extreme political views and you're having trouble unseeing them.
If I have that right 👆 it's something we want to do but we're early and small and picking our battles, and wanting to make sure we really understand the problem.
This is what G+ was trying to do with Collections, and I'm fine with that. My hip hop channel on G+ was more than twice the size of any of my other collections. As long as someone can follow all of my technology posts, for instance, that functionality works for me.
So, if I like drone photography (which I do) and I start reading and following content in that area - how do I filter out the content from people that I don't like or aren't delivering what I consider to be credible content?
Hey David, we do have mute and ignore in place now.
It's almost like you were reading the post I was writing ten minutes ago:
"For example, we don't care about how many users have registered. But we do care about how we get great conversations that get lots of views."
And how do your investors feel about that?
hmmm. If I go to a post and click the drop down menu icon next to it I just see flag. Am I doing mute wrong at the post level?
Thomas, you are likely looking at the dropdown at the top level of a conversation, which only has "Flag" as an option. The owner of a conversation has more options there. However, right next to every post there is a dropdown that lets you "mute" or "ignore" a person. Muting will prevent that person from posting in your conversations, but it won't affect conversations you don't own. Ignoring a person will collapse their post in any conversation on Cake whether you own it or not.
I agree with you about the arc of social networks but I feel that the case of Google+ is a little different. Though more difficult to find, I think the organic conversations continued and may continue until the lights go off. I think that G+ ultimately failed because of lack of a cohesive - and simple - vision of what it should be.
I don't want to move this conversation to the fate of G+ (there are other posts for that) but the cautionary tale of G+ is you need to two ingredients to make an interest-based network successful: Vision and love. That's the burden for Cake to deal with now.
Ah, ok, it might be useful to be able to mute/hide an entire thread. If I know for sure 100% that a thread is of no interest to me, every time I have to see it again on the site it becomes noise.
Thank you for the insight and honesty - the eight problems resonate. I must confess that, although I've been looking around here for a few days, I hadn't realized the full benefit/potential of panel discussions. Now that I do, I'm looking forward to sitting back and watching experts do their thing!
What makes you think that bad actors will not create Panel discussions which are filled with other bad actors? Anyone, as far as I know can create a panel discussion, so what is to prevent toxic panels?
hopefully we will get tools at some point to mute panels and conversations like those. :)
Good question, Chris. 🙂 The answer I give them is eventually we'll have to run ads (interest-based because we know what topic everyone is talking about, we don't need to do surveillance capitalism). Aside from interest the ads depend on audience size. So if some famous celebs do a panel that gets a zillion reads, I think it's good for advertisers in the same way it's good for investors when a YouTube gets a zillion views. Maybe better than a million people chatting about something mundane.
(We also hope to do the equivalent of YouTube Red subscription in the future to let people opt out of ads.)
That sounds like a very good idea; maybe add Hide as an option like we have Follow now so you never have to see it again. Our engineering team are probably wincing thinking about making that work well at scale.
Ugh. I hate to be redundant, but the lack of threaded conversations here is really challenging. I would think that a comment parent ID would be trivial, but not having any idea what your architecture looks like, I'm sure it's probably more complicated than I think.
The interest based ads model has worked reasonably well for Reddit, but it wasn't enough to fully solve their revenue issues. Reddit Gold and Cake...Frosting?..seem like a reasonable premium model, depending on how efficient your stack is. If your interface starts lagging hard (and even the Plus had issues with this), your burn rate is going to increase faster than planned just on infrastructure scaling.
I don't think "Opt out of ads" is sufficient on its own for Premium, though. My uBlock Origins is doing an adequate job of that, so what incentive do I have to pay? Brand loyalty factors into the equation (and I have been a Reddit Gold user since it began), but it was a ten year cycle for Reddit to build up to the point of making any serious money off of Gold. Perhaps a more evolved Freemium model?
BTW, I'm not just a gadfly; I'm really happy to be here, and glad I saw Robert's post on the Plus. I've experimented with every new platform that's come along, but they were all deeply flawed, and certainly none of them have the pedigree you bring to the table. I sincerely hope this platform takes off.
love the option to pay a subscription and opt out of ads when they show up. :)
I hate to be redundant, but the lack of threaded conversations here is really challenging. I would think that a comment parent ID would be trivial, but not having any idea what your architecture looks like, I'm sure it's probably more complicated than I think.
It's a design decision, not a technical limitation, but I hear you. Behind the scenes we do have all the data necessary to support a threaded view, and we've talked about offering both threaded and flat views as options. Just takes time. 🙂
I am nothing if not patient with emerging platforms, especially when the principals acknowledge feedback. Brian Rose was really good about that over at Plus, which is part of what led to Google Photos.
I agree with the flat view design decision. I’ve always preferred that but to each their own.
There are plenty of places where that works well, no question. But from a pure usability perspective, when I'm replying to a comment high up in the thread, and bounced to the bottom of the thread to do so, it's a ton of extra work to keep scrolling back up to find a non-highlighted non-quoted comment to pull sections out to reply to. If the comment you are replying to was fully quoted in your reply by default, or if there was a UI element which showed the comment you were replying to while composing, the experience would have much less usability friction.
Also, this is only going to get worse when we start having 1000 comment threads.
On the arc of social networks, I have a huge motorcycle forum and motorcyclists are pretty frisky. Over the years as we've created subforums we watched the arc go from hard-to-get-traction ==> positive network effects ==> awesomeness ==> it's getting hard to keep up and it's getting noisy in here ==> it's getting impersonal, who are these n00bs? ==> oh look, here come the spammers and trolls.
We would then slice the subforum into pieces. We'd get a few howls but 9 times out of 10 it would restore the pieces back to positive network effects. Some would stay small enough, like Old's Cool, a subforum for restoring vintage bikes, that it has stayed great for years.
Our hope was that by enabling the topic structure we have, people could create very specific topics, like Ural Patrol sidecars, to keep the signal to noise high.