I strongly resonate with the part about Circles being an active-maintenance-effort that even 10 years ago, when I actually had time, was more effort than I could be bothered with.

I think an interesting question is, what would it take for google to create a social network that would succeed nowadays?

They have quite the reputation [1]. If Google suddenly came out with an all-other-social-network-killer, and successfully killed all the other social networks, I'd still expect them to shut it down three years later.

... which unfortunately puts them in the position that it'll be incredibly hard for them to put out a true market-segment killer for anything, nowadays. Can't take over a market segment if no-one trusts you can be relied on, and therefore doesn't move to your product.

I say this even though I met my wife on Orkut, one of Google's previous social networks that, sure enough, got shut down.

[1] https://killedbygoogle.com/

RE: the circle maintenance effort, I’m finding that people are using group chats (in messenger, WhatsApp, etc) as a solution to the pain points Circles set out to solve.

For example, I’ve now been added to a chat with all my extended family, one for my local community group (and subgroups), and a couple friend ones from college. And now when I have something to share I pick the relevant scope just like Circles, and finding my contacts are more open to sharing casual thoughts vs standard posting.

I have a hunch that it works because the moderation/curation becomes a group effort instead of just a maintenance activity asked of me alone, but it was an interesting realization when I realized what was happening

Not only was it individual effort, it just felt so pointless. In group messages (Facebook/WhatsApp) you're effectively/literally inviting someone into a group conversation like a private forum amongst the invitees.. In g+ circles, all you were doing was labelling your friends and even naming the labels could be tricky ("friends" circle .. if you aren't in it, what does that say?)

Same reason I don’t start watching any new series on Netflix anymore. They’re building a similar reputation for themselves, cancelling shows before the story can be completed.

If it’s a miniseries (e.g. Queen’s Gambit), then I’ll try it out. Otherwise I won’t watch something if the rug’s going to be yanked out.

Which of course adds to the likelihood of that happening, if others take the same wait-and-see approach.

Shows being cancelled earlier than fans would like is hardly a phenomenon exclusive to Netflix.

While this is off topic I blame the writers/producers.

They get money to produce 1 season. If they end that season with a cliffhanger there is a larger chance that viewership demands a season 2, than if it ends cleanly.

I feel like the writers are taking the viewers hostage, so to speak.

I guess you could say that Netflix should demand season 1 to stand on its own. Or at least be clear if a given season ends cleanly.

What are the series on Netflix that you feel were cancelled too soon? The only show I was disappointed to hear was 'cancelled' was Mindhunter, and I just discovered it's actually going to be coming back!

https://gizmostory.com/mindhunter-season-3-release-date-rumo...

EDIT: In response to the comment below me, I am sad to learn Glow isn't coming back.

Mindhunter was the biggest letdown when I found out they axed it. I hope they do bring it back.

The OA and Santa Clarita Diet are a couple others off the top of my head.

There’s a lot more, but I’m already not watching them until they’ve done 3 seasons. (Unless of course they’re intended from the outset to be just 1 or 2. That’s different)

I was unreasonably sad when The OA was cancelled.

If you think about it, The OA wasn't cancelled, was... _finished_, since they ended up in current reality.

Altough I'm sad that the show ended, I'm pretty happy on _how_ it ended.

A quick internet search finds lots of examples. Here are 27: https://www.businessinsider.com/best-canceled-netflix-shows-...

I, too, largely have stopped watching Netflix shows with one season for exactly this reason

I looked through this list and at least Lucifer and Sense8 I felt they completed the story they wanted to tell. Especially with Lucifer I don’t think they can extend the series longer with keeping the style but without making ridiciloud new additions to not make it boring and repetetive.

Films don’t go on forever, neither do series. I don’t think putting finished series together with outright cancelled series is fair, and thus I have no idea what all the other items in the list mean.

Lucifer is an odd one. It's demonstrably terrible, technically, but it does have a cult following.

But it's also got its 6th, final, season coming in 2022..

Besides what others said: they canceled BoJack Horseman when it had 95% audience approval rating in its final season (!!!), and had approval ratings above 80% do the entirety of its run.

The absolute dumpster garbage that is Paradise PD that can barely scrape past 50% rating is now in its fourth season or something.

It is a well known in TV the higher the approval rating, the smaller the audience. Hence the phenomenon of well-liked shows being cancelled, networks like audiences over quality.

And it was only cancelled because Fincher can't figure out how to make a basic no-vfx cop show for less than $100M.

(Don't get me wrong - Mindhunter is amazing TV. I just don't see what justifies the cost).

See, i watched 3-4 episodes and I just don't get it at all. I don't care about a single character and I am not sure any of the episodic plots are interesting in and of themselves.

Like, I am missing something there.

Even before Netflix, I watched only first season of any popular TV show. Lost interest after that.

YouTube is already Google's killer social network. Over a billion users a month.

This I think is the observation that many people miss. YouTube, while not identical to Facebook in its approach, is a massive social network, and massively successful, with a lot more goodwill than Facebook gets.

I think the good will comes from YouTube doing something useful for the user with the collected info (video recommendations).

And without ad-blocker circumvention.

Is Youtube actually a social network or a broadcast medium with a very narrow feedback medium? All the big channels I know about (only briefly) have forums etc external to youtube.

In my view, a social criteria needs to have the following criteria:

- anyone can make a post (video, text, image, etc), either to an individual (instant messaging) or to a group (public or selected people)

- people can reply to other people's posts

- communities form automatically from the above two interactions

Youtube is more a para-social than a social network. Social networks are all about interaction on equal terms, but on youtube the focus is strongly shifted to creating videos and let them concumed by their peers. Proper interaction hardly happens there. And didn't they even remove the means for private communcation some years ago? So it's all just public now.

The only somewhat social interaction happens between the content creators, not between any user of the site. And those creators seem to care more about beefs and making money.

I see creators engaging with comment writers all the time, as well as commenters engaging with each other on a community level, particularly in more niche communities

Indeed, but that is on the level of any other comment-section. You can't follow other commenters and subscribe to their comments. Videos and comment-section are different worlds, with the comment-world having network at all on the technical level.

Social vs. parasocial is simply a matter of the scale of any one node.

An individual (or organisation) can only maintain so may relationships. Dunbar's Number is a pretty good proxy. Ultimately, what each relationship is competing for is time, and that is a nonfungible resource. Ultimately, our relationships are defined largely by arity, the number of connections between parties and communications, the n:m relationships: 1:1, n:1, 1:m, n:m.

The vital groups I've been involved with have been 1) highly curated and 2) about 40--50 people, of whom maybe 10--20 were highly active. (One such group was a privat G+ community. Whilst it lasted it was excellent. I ultimately opted to kill it off due to concerns with the platform.)

A person, or a group, with more than a few hundred followers, can at best devote only a few minutes or seconds to each.

Over an 80 year life, the time you can devote to individual members of a group (assuming you sleep 8 hours/day) is roughly:

  10:  5 years
  100:  6 months
  1,000: 20 days
  10,000: 47 hours
  100,000: 5 hours
  1,000,000: 28 minutes
  10 million: 3 minutes
  100 million: 17 seconds
  1 billion: 2 seconds
  10 billion:  1/4 second

Google+ existed for slightly less than 8 years, and had (generously) 3.3 billion profiles. Each user could interact with each of the others for ... 0.05 seconds.

If we rely on a more realistic estimate of about 100 million actively-engaged profiles, that engagement increases considerably to ... 1 2/3 seconds.

In a parasocial relationship the fandom devotes a significant portion of its life to the star (often minutes or hours per day). The star ... cannot devote much time to individual fans.

The largest followings on G+ tended to be around the 10 million mark, allowing about 17 seconds of star attention per fan if the star devoted every waking hour to such relationships.

My more modest 2,000 or so followers might have received a days' attention from me, were I particularly obsessed with such engagement.

In practice, most of my engagement was with 20--100 or so people. Several of those I've continued to stay in touch with.

We seem to be under the delusion that technology allows us to scale intimacy. That's a fundamental impossibility, as true intimacy (measured in both time and distance) simply cannot be mediated. The problem isn't just the arity mentioned above, but context, expectations, understandings, intentions, and more. Intermediated communications quickly gets quite complex. Intimacy does not scale.

https://joindiaspora.com/posts/4bdf7470aa0301394c70002590d8e...

> Social vs. parasocial is simply a matter of the scale of any one node.

Not really. But parasocial comes more natural with scale. The main difference is the nature of relationship and communication. You can have a parasocial relationship even in a 1:1, and a social realtionship in a 1:billion.

The defining point is the type of power a relationship has. In parasocial it's one-sided, parasitic. While social is multisides, symbiotic. Though of course there is nor purely social or parasocial relationship in any group, there are some which are more like one or the other.

> A person, or a group, with more than a few hundred followers, can at best devote only a few minutes or seconds to each.

Social does not mean you must share personal connections with each individual. The crowd can also interact anonymous and indirect. The crowd can be social or parasocial, depending for whom they act and how much power it's head has.

Yes, it's possible to have low-arity parasocial relationships.

No, it's not possible to have high-arity social relationships.

> Build around people problems, not company problems

This is so hard to fight for. I worked on a feature for our product where we leverage a UI component initially built to show ML recommendations, but to surface hard coded recommendations made by our partner's software, not coming from AI but hard coded. These recommendations are contextually sound because the devs _know_ what is possible given the context, which for all intents and purposes makes them much, much more relevant than anything that would come from an ML model. But for the life of them the other team will not let us use this UI because they're afraid that it might tarnish the opinion people have of their AI. So we're in this situation where the user would greatly benefit from a consistent UI, the partners as well, but because that team is afraid it will dilute their baby, we can't do it.

Had similar examples where 3 months in designinga feature it becomes clear that it doesn't fit with users, that a simpler solution would be better, but designers and PMs will take a hard stance that it's how we planned it and now is too late to change it even if dev hasn't even started.

Between sunk cost and babies, it's hard to forget our own interest in favor of our users.

> sunk cost and babies

It’s not that they (middle managers) are stupid. It’s political. KPIs and objectives and individual careers are misaligned or work against what is good for the user and ultimately, good for the company. Blame the CxOs.

Yeah that's absolutely true, things get done not because it's the right thing but because it will reflect or benefit people. If you create wrong incentives you'll get the wrong behaviors in return.

Conway's Law in action. You have my sympathy and condolences. https://en.wikipedia.org/wiki/Conway's_law

FWIW, Tesla seems to be more willing to shake the ant farm. The tear downs and comparisons with other EVs are instructive. Tesla's octovalve, heat pumps, gigacasting, and I'm sure a dozen other things, just wouldn't happen (didn't happen) elsewhere. One blocker was functional silos.

Maybe it takes a tyrant like Elon Musk to dare the things Deming, Drucker, Goldratt, and others were proscribing.

Or external, existential threat. Like Churchill said about America: They always do the right thing, after first trying everything else.

In my experience, the only times I've successfully busted silos, the circumstances were dire and everyone else had given up. Other times were much less successful.

>I’m sure there was a ton of legacy to deal with here, but you have to wonder about what might have been if an identity layer had simply been applied across all these products and they were invested in.

Google kind of did this though, and everyone hated it. Remember on YouTube, the mess they made when they introduced Google+. Maybe the author means "if a single identity had existed from the beginning..."

I want the best of de-coupling and coupling of services, where you only need one identity to manage a variety of services, but each service is isolated from one another. This comes in two forms: identity and repercussions. In terms of identity, I want the ability to present myself differently depending on the service their using. I don't want my real name in say a game lobby where harassment happens more regularly. In terms of repercussions, I don't want a TOS violation on one connected service to influence the other services. I hear a lot horror stories of where malicious intent from a 3rd party on some service causes the closure of all coupled services and you're left dry.

I think Microsoft does a reasonable job at this. If you violate the TOS of Office 365, you may lose access to that, but the rest of your account stays usable.

Didn't Yahoo have something vaguely like this? I don't remember the details too well, just some vague memory of Crockford bemoaning how well Yahoo services were segregated from each other compared to Google.

At least the old pre-Verizon, non-Japanese Yahoo. Japanese Yahoo is very integrated 'till today (banned at Auctions? There's a big chance of your account being shuttered down.) and Verizon's Yahoo has gone down to mail and news (okay, there's Yahoo Finanace (the products, not their newsroom), but I don't think that most people are using it now, and Yahoo Answers being gone is one of the most prominent examples).

That's a good take. I agree with presentation, and I think Google fixed this on YouTube to some extent. However, they're failing at the repercussions aspect.

Yup, I lost access to bing rewards on my main account but I can still use it for everything else. (Including searching, I just don’t get points for it).

They did a terrible job of it. They broke all existing comment trees and pushed google+ really hard and automated posting. I seem to remember that commenting on YouTube would then make a post on google+ with your real name and likely notified your contacts about it.

They clearly didn't want to be left stuck with maintaining multiple authentication systems (legacy ones and G+), so they thought they'd eat shit for a while but results would be worth it. But they executed so forcefully, that the backlash eventually forced them to backtrack on the most visible properties. This said, now every product they create or acquire gets wired to a single authentication system, so I guess they did make some progress, even though it's a "Google account" rather than a "G+ account".

I learnt that you shouldnt bother investing any time into any new Google service after 2010.

Google drive and google photos are after that time and they both seem to work really well and be core products.

Google Drive is mainly a response to Apple's iCloud, Microsoft's then-SkyDrive, and Dropbox. It was initially an integration into Gmail, and became a fully-fledged separate product later.

Photos is the spiritual successor to Picasa, integrated into the web. It was a response and a one-up to (again) iCloud and SkyDrive which both have phot syncronisation, which is something that is missing in Android. It was initially integrated into Google Drive, and became a fully-fledged separate product just now, when the integration has been cut.

So Google's next successful product would be

1) When Apple and Microsoft has bothered to be the first, and

2) whatever is connected to Google Photos.

(btw, since Stadia has became a meme, it partially fits 1 (xCloud) but sure it isn't connected to Photos in any way)

Stadia won't be able to compete with GeForce Now anyway.

> And to put the elephant in the room to bed, Google didn’t need real names, they simply needed one common way to refer to individual people.

I don't think this is sufficient treatment of what was, at the time, the biggest objection to Google+ that I heard. The hamhanded forced real name thing killed Google+ for virtually everyone I knew on there, less for the actual concern about using your real name than for the attitude it displayed towards users' privacy.

I know I was well on the way to moving completely over to Google+ at the time they announced that change, and dropped it entirely almost immediately thereafter.

Why didn't they implement basic boolean operators over circles and people?

Post (@friends !bob) "let's surprise Bob for his birthday" is about as normal as it gets

> Why didn't they implement basic boolean operators over circles and people?

> Post (@friends !bob) "let's surprise Bob for his birthday" is about as normal as it gets

Because while I would love that UI, there would be 1% of the population on their site.

Look at how much they had to simplify search syntax to appeal to the wider market.

You can always represent this with a widget and a little form.

But from my experience, people who quit and say “I don’t understand” in front of the login screen of Windows, have no problems setting up an OnlyFans account with 3 levels of verification and linked with a bank account.

> But from my experience, people who quit and say “I don’t understand” in front of the login screen of Windows, have no problems setting up an OnlyFans account with 3 levels of verification and linked with a bank account.

So true that it is almost sad.

Setting up an OF account doesn't require choices, it requires steps that the UI designers are very motivated to have you step through.

That's why they don't give up

Syntactic sugar it into natural language, I just wanted grouping, braces to coerce precedence with "and", "or" & "not".

I even said so at the time! (Back when we naively believed Google wanted feedback)

It's not a syntactic sugar problem. Most people don't think like that.

My 40 years of experience in computer science thinks you're wrong. Google agreed with you, so, there's that.

> My 40 years of experience in computer science thinks you're wrong.

Why would 40 years of computer science experience be relevant to predicting how most people behave? If you had said psychology or something, I can see it being relevant.

You know google support boolean terms in their search products, right?

"most people don't think like that" is an odd statement. How do you know how most people think? You think most people don't understand how to ideate "all my friends except bob" to invite people to bob's surprise birthday party?

Thats syntactic sugar, right there. You don't need a degree in psychology to get this. You need a degree in computer science to know what syntactic sugar is.

> You know google support boolean terms in their search products, right?

I know they used to. I know people claim they still do. If they do, they seem to only treat it as a suggestion (like searching for a significantly less popular by very similar word) and ignore me when I tell them to exclude specific terms.

> How do you know how most people think?

Okay, I'm not going to play the game where I have to keep supplying evidence and you keep denying it. What do you want. Google agreed with me, you have no training in anticipating how people think and the rarity of people capable of being programmers are sufficient for me to say it's time for you to start supplying proof of your blatant assertions.

> You think most people don't understand how to ideate "all my friends except bob" to invite people to bob's surprise birthday party?

Correct. And certainly even if they can it's only because that example is particularly simple (it's the same as always only without Bob). It doesn't solve 99% of use cases.

I'm not disputing that syntactical sugar is nice, I'm saying people are not capable of the complex operations to make the system you are positing work the majority of time for the majority of people, they will either quit in frustration or send their drunken party pictures to their boss or mother.

Maybe the failure with google plus did them a favor: so instead of owning another social network they focused on analysing all the private data they already vacuumed up and pushed machine learning into it every corner of the system, maybe this preceived lack of private data forced them to be more thorough at analysing what they already had and forced them to focus on android as source of data.

what is Vic Gundotra doing these days? I loved his Google I/O presentations .

Seems he's taken a hard turn to the alt-right:

https://nitter.eu/GundotraVic

Facebook's success was all about network effects. It's no coincidence that it started out at Harvard. That was the root of the network where the most influential people of the next generation would come from.