Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

theverge.com

Post reply on HN
Print view View original

471–480 of 585 posts

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#471

Post by shagie »

N_A_T_E wrote: post #241

I'm assuming 90% of these API requests are reads. Is there a reason Apollo couldn't develop an intermediate proxy server and cache all the responses for a short period of time? I assume that would cut the API usage significantly and potentially the cost.

They are mostly reads. They are done every few seconds for each user to get the mailbox status for push notifications for monthly subscribers.

https://youtu.be/Ypwgu1BpaO0?t=1772

> I guess there's an analogy um the way Reddit notifications work just for your inbox like you got a message or something um they work in so far as if I the developer of the app want to um say make sure that you get that notification within 10 seconds I have to be checking Reddit every 10 seconds to go like is there anything new is there anything new? is there anything new? is there anything new? okay. There is okay I'll tell is there anything new and then just repeating that at nauseam so you can imagine if you get a message once a week I'm checking every 10 seconds and then once during that whole week I get that message and then I can send it to you um so 99.99 of those API calls were wasted so we've talked to Reddit like that my friend who works on my server um and myself and I've said like what would be so much better is if we could just kind of keep like a port open with Reddit and say like you just tell us when there's a notification ready and we'll beam it off we don't have to bug you all the time and it's logical right and that's how a lot of services do it it's like an event-based API and um that's just not something reddit's ever uh given us

Note that it does a request every 10 seconds for each user. At the API rate that would be about $0.25 every 3 hours (for each user) to support it.

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#472

Post by stetrain »

evandale wrote: post #467
stetrain wrote: post #411

Earlier quoted context omitted.

One phenomenon that I think drives this is that the engaged users (ie commenters) are usually a small subset of the total user base that upvotes things. The engaged users care about the topic of the sub and tend to interact with it directly, while the larger audience is subscribed but mostly upvotes things from their main feed without caring which sub it is from, or visiting the sub itself to see what all has been po…

> I have seen many cases where the “just let the votes sort it out” method leads to things being upvoted, presumably by those users scrolling their main feed who don’t even notice what subreddit it is from, and then comment sections full of “who is upvoting this junk” “this is the third time this has been posted this week” and “mods can we please do something about all of the X posts?” Oh come on, you can't be seriou…

> Oh come on, you can't be serious. These are different groups of people. The same users who upvote things are not also complaining about what is being upvoted.

Yes, that was the entire point of my comment.

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#473

Post by actionablefiber »

mbStavola wrote: post #9

I can't say I've missed Reddit too much after I stopped using it. Twitter was a bit harder, but Mastodon filled that niche quite nicely. However, I don't even feel like checking out Lemmy or Kbin as a Reddit replacement, making me question how useful Reddit is for me personally. I wonder if anyone else is feeling the same during the blackout.

I moved Apollo to my last page of apps instead of my first. It taught me a lot, very quickly, about how compulsively I was using it, and how little I missed it after the unconscious act of opening it up left my muscle memory.

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#474

Post by sunflowerfly »

adrr wrote: post #435
protastus wrote: post #210

Earlier quoted context omitted.

> Not really mentioned was all this was changed out the blue with less than 30 days to comply. The interview touches on it but it’s insanity that a tech company think it’s reasonable for indie devs to operate under that constraint. It's unreasonable for a company of any size. The entire storyline shows Reddit sabotaging the process. Which seems idiotic, because they're throwing away a vibrant ecosystem that is #1 in…

Isnt that what the CEO is trying to do with charging for API access? Twitter did the same thing. Apollo dev could just raise rates to $3 a month and he makes money and the user in return get a apollo which offers an ad free browsing experience.

A point that has gotten lost in the discussion; The CEO is saying the "average" user would only cost $1 a month and therefore the pricing is fair. But users of apps like Apollo are not average, they are super users (and include a high percentage of mods). The Apollo app developer says he would have to pay $2.50 a month per user.

https://www.reddit.com/r/apolloapp/comments/13ws4w3/had_a_ca...

If he was able to pass through $1 a month to users and given a reasonable time to do so, Apollo would likely still be alive. Or better yet, require a $10-12 a year subscription on each Reddit user account, paid directly to Reddit, to use 3rd party apps.

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#475

Post by shagie »

mishagale wrote: post #68
causi wrote: post #8

I would like a breakdown of Reddit's actual costs vs their proposed fees. Why isn't Reddit profitable? Is there a way to make Reddit profitable without charging for API access?

Based on their published user numbers and revenue, Selig estimated that reddit's revenue per user is about $0.12. They are asking for $0.24 per 1,000 API calls. Apollo averages ~345 calls per-user per-day (although some users are much higher), but Reddit claims that other apps are "more efficient" and only use about 100 calls/day. If we assume 100 calls/day/user, then that's 36,500 calls/year/user, and a user is wort…

> Apollo averages ~345 calls per-user per-day

The mobile front end does about 345 calls/user/day

The push notification server does 8640 calls/user/day (one call every 10 seconds).

In https://youtu.be/Ypwgu1BpaO0?t=1772 he describes how that works.

> I guess there's an analogy um the way Reddit notifications work just for your inbox like you got a message or something um they work in so far as if I the developer of the app want to um say make sure that you get that notification within 10 seconds I have to be checking Reddit every 10 seconds to go like is there anything new is there anything new? is there anything new? is there anything new? okay. There is okay I'll tell is there anything new and then just repeating that at nauseam so you can imagine if you get a message once a week I'm checking every 10 seconds and then once during that whole week I get that message and then I can send it to you um so 99.99 of those API calls were wasted so we've talked to Reddit like that my friend who works on my server um and myself and I've said like what would be so much better is if we could just kind of keep like a port open with Reddit and say like you just tell us when there's a notification ready and we'll beam it off we don't have to bug you all the time and it's logical right and that's how a lot of services do it it's like an event-based API and um that's just not something reddit's ever uh given us

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#476

Post by kenhwang »

duped wrote: post #469
kenhwang wrote: post #448

Earlier quoted context omitted.

There's a lot of ad fraud in the industry, there's a lot of money in advertising. Everyone is incentivized to say they showed an ad but not actually show the ad. There's a lot of very sophisticated ad fraud operations and the tracking is designed to be difficult to defeat. We've seen app developers try to render ads in hidden iframes or muted behind/under content or on server farms with headless browsers or click scr…

> Everyone is incentivized to say they showed an ad but not actually show the ad. People don't want to see ads because ads are intrusive. The solution to this problem that the ad industry sees is to make ads more intrusive. So it's kind of a positive feedback loop of user-hostile garbage forced upon them, and no wonder that people want to make ad-free experiences.

I'd say there's two kinds of publishers.

The "good" publishers will simply charge more per ad and do more hidden proof of trustworthiness, such as being more open to code audits/transparency, more user tolerable ad practices/product deisgn to minimize bad behavior/increase tolerability, or better ad targeting to simply demonstrate more effective conversions. Think Facebook (the tradeoff here is they know a lot more personal information about you to be able to do this).

Unfortunately, most ad publishers fall into the "bad" camp where they know they won't receive payout for some portion of their ads, so they increase the amount of ads shown to make up for the shrink or simply show lower paying ads that have more lenient verification requirements. This is more what Google does/encourages.

My experience is advertisers would love to show less ads with higher conversions, because it makes their lives way easier. But they can only buy what's offered, and there's a lot of cheap user hostile publishers desperate for ads (even most of the ad strategy suggestions in this thread are pushing for more ads) and very very few high quality ad publishers.

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#477

Post by actionablefiber »

tyg13 wrote: post #457
penguin_booze wrote: post #212

Maybe I'm missing something: why can't Apollo--or any other app--pass the API burden to the user? I get that this means the user experience via the app would effetively turn out to be paid service. But then, to match, Reddit could have also relaxed their policy to the effect that, up to n API access would be free; beyond that, Reddit would charge the user. Microsoft, for example, already do this for many of their Azu…

This was something that the Apollo dev (and a few other apps) were interested in, but the timing was just too sudden to make work. Specifically for Apollo, they had been selling year-long subscriptions for $12/year. With the API pricing change happening on July 1, that meant Apollo would have to continue to offer up to 11 months of the same service, at the same price (since users already paid for it) while operating…

Seems it should be possible to shut down, refund, and relaunch under a new pricing model? Does Apple forbid that?

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#478

Post by evandale »

shagie wrote: post #444
ryanwaggoner wrote: post #438

Earlier quoted context omitted.

I see you making this claim in multiple comments that he's making 600 API requests per user per hour to check for notifications, but this doesn't make sense to me. As you said, 100k users would result in 60 million API requests per hour , which would be about 43 billion API requests per month. I've seen pricing of $0.24 per 1k calls, so that would be $10 million per month, not $20 million per year. And that's just fo…

It's possible. He could have also changed the rate at which it polls. ~1B calls a day isn't out of the realm of possibility. https://www.reddit.com/r/redditdev/comments/13wsiks/comment/... > On March 14th, Apollo made nearly 1 billion requests against our API in a single day, triggered in part by our system outage. After the outage, Apollo started making 53% fewer calls per day. If the app can operate with half the d…

Linking to Reddit admin posts isn't going to do you any favours. The Apollo dev has addressed, acknowledged, apologized, and fixed that March 14th problem and it's all detailed in his post [1]. He has given everyone the receipts, phone recordings, source code, and emails detailing his conversations with Reddit. There is no reason that I've seen to not take his word at face value or think he's trying to mislead people or make Reddit look bad for bogus reasons.

The Reddit admins have done nothing but lie and try to bend the truth and I would never trust their word after this debacle. Christian has far more credibility than the Reddit admins do at this point and their "well Apollo reduced usage by 50% one time with a 24 hour turnaround so that means their app sucks and surely they can do it again if they tried a little bit!" argument is, for lack of a better word, bullshit.

That isn't how software works, Reddit knows software doesn't work that way, and it's such a ridiculous conclusion to come to that I don't even know if it's a fallacy that has a name but the Reddit admins are definitely not arguing in good faith on this.

[1] https://old.reddit.com/r/apolloapp/comments/144f6xm/apollo_w...

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#479

Post by adventured »

redmerchant2 wrote: post #7

They buried the lede. The article suggests the shutdown was because of free API to paid API. Anyone uninformed who just click off at that point. Even when they ask about costs, it was just saying it’s “ But the issue is two fold. (1) all the devs support a paid API, but the rate is absurd according to basic math. It would charge devs over 20x what a user actually cost/were worth. Devs were willing to take on a higher…

This isn't about API costs. It was changed so rapidly due to ChatGPT 4 and how it shocked the industry. They believe time is essential, in regards to pricing up the data (GPT has indicated to the Reddit company that it has massively undervalued its text content).

API pricing/use is the excuse to lock down their content asap. They see a goldmine under their own feet that others are extracting value from and so far they have not.

This is all about LLMs and perceived value related to that.

Re: Apollo’s Christian Selig explains his fight with Reddit – and why users revolted

#480

Post by sunflowerfly »

tedivm wrote: post #60
0zemp3c wrote: post #30

Earlier quoted context omitted.

99.99% of reddit users have no idea who spez is and probably wouldn't care if they learned about his shenanigans 99% of reddit users are aware that moderators do some nice volunteer work but wouldn't cross the street to save them its only Extremely Online people who care about this Insider Baseball stuff

Sure, but without those moderators there's not much of a site. I don't spent a lot of time thinking about trash collection, but if all the garbage on my street stopped being collected I'd absolutely notice.

I hit Apollo yesterday out of muscle memory and clicked on one of my favorite subs. The posts were simply terrible. If that was the daily experience I would not ever read that sub.
Post reply on HN