Earlier quoted context omitted.
What I think the grandparent is getting at: You can measure the direct effect of a change now on something like conversions. But you can't measure the second order effects: things like trust from your users, or the effects on community quality and composition, etc. This is a good part of why enshittification happens: lots of changes with immediate "good" impact that can be measured quantifiably, but there's also read…
Generally you'll have a quarterly holdout to measure total impacts but I agree that yearly would be better.
The Tyranny of the Marginal User
701–710 of 840 posts
Re: The Tyranny of the Marginal User
#702I did?
I mean, you can meet a significant other or friend in any kind of place. I kind of get what the author is trying to say but it's ill phrased.
Re: The Tyranny of the Marginal User
#703This is a hilarious read but I think the author is too optimistic about the state of humanity. Marl isn't the "marginal" user, Marl is the "average" user. If the average user actually cared about deep and meaningful content, then any A/B test that throws her under the bus in order to please Marl will show bad data, and the proposed change would be killed. Yes, the author tries to hand-wave this away as "product is st…
A/B tests, as they are run by current software companies, are inherently flawed. I have never, in my entire career, ever heard of an A/B test that ran for a year, let alone 3-5 years. That’s where the true power of statistics comes alive, and nobody is financially incentivized to even consider that fact.
https://www.zumsteg.net/2022/07/05/unchecked-ab-testing-dest...
Re: The Tyranny of the Marginal User
#704I worked at OkCupid from 2013-2017 and totally resonate with the author that mid-2010s OkCupid was a really special product, and that it took a steep decline as the decade went on. It's not entirely fair to say that the Match acquisition immediately caused that decline; I started a couple years after Match got the company in its hands, and only two of the original founders were still focused on OkCupid full time. But…
I met my wife on OkCupid. The original format attracted a much smarter and more worldly crowd of women, to put it bluntly, than the other services. I exited the dating game before Tinder, but if OkCupid lost that quirky, artsy, college educated crowd in the chase to compete, that's a real shame.
I had another OKC date from that era where the woman had a very high IQ, her father was a prolific author and her late 20’s brother was VP of ask.com at the time.
I had just assumed it was the Silicon Valley bubble clientele (and still just might’ve been). But only 3 years later I recall younger male coworkers describing how much dating apps declined and how terrible the dating scene had become with “swipe culture”.
Re: The Tyranny of the Marginal User
#705Earlier quoted context omitted.
Nope. It's not specific to their height and obviously not specific to their race. The same way men on average prefer big boobs, women on average will prefer taller men. Both can't really be changed by the person but exist as attractiveness scores against a person. It matters to people who it matters to. Not many in the grand scheme and definitely not related to race.
I have literally not once in my entire life heard a woman in real life say 'oh he's so tall, that's so attractive'. That's only an internet thing. Or an outside-Western Europe thing, maybe.
I’m glad you haven’t had that experience. Doesn’t discount that it’s a real thing.
Re: The Tyranny of the Marginal User
#706Earlier quoted context omitted.
If you listen to Frances Haugen's testimony, you see that it is very much deliberate. There are probably thousands of psychology grads that went to school, hoping to help people, that are, instead, designing dark patterns. But they are probably making a lot more money than they would, helping people. Evil pays cash.
It's as deliberate as a cow eating grass. The cow deliberately eats the grass but because of the game theory involved (lack of information, lack of coordination, incentives etc.), the pasture disappears. The point of my comment (and the article's, I think) was that companies want the revenue, but don't deliberately want to harm us.
Re: The Tyranny of the Marginal User
#707Re: The Tyranny of the Marginal User
#708Earlier quoted context omitted.
This is basically the foundation of modern, online dating -- aka red pill. Roughly, the top 20% of men get 80% of the attention from women. (I'm pretty sure that figure comes from Tinder data.) It makes sense from an evolutionary perspective, but is brutal in the real world. What do you do if you are average (or less) in looks and income (potential)? Prepare for a lonely existance.
[flagged]
Re: The Tyranny of the Marginal User
#709Earlier quoted context omitted.
Sure, just make something that has a barrier to entry that will filter out non-nerds. From what I can tell Ham radio is basically just a chatroom with an entrance exam, for instance. Anything done in a constructed language like Esperanto would be another filter. There's nothing stopping anyone from making a website duplicating the original OKCupid method. It's just that it requires someone willing to say "no" to the…
Can we use separate root DNS server as technical barrier to entry? Or NNTP protocol. Or client side SSL certificate, which must be won in a complex game. There are plenty of options to separate smart people from regular crowd. IMHO, a separate DNS server, which unlocks alternative Internet, is the easiest way to move away from regular people, because it's a bit complicated to switch DNS on a mobile device.
Re: The Tyranny of the Marginal User
#710Earlier quoted context omitted.
I think the changes to Reddit suggest that we (the ones complaining about Reddit) are but a small minority. I really thought Reddit would revert their API changes after seeing the community response, but then... nothing happened. This was the event that made me realize how far I am disconnected from its average user. OP makes it seem like Reddit's users lose something when Reddit panders to a Marl. But I've observed…
I can't say what percentage of Reddit users cared, but I can definitely say that the majority of Reddit participants—the people actually posting the content Reddit is trying to sell—have left. Reddit activity has dropped off a cliff: https://subredditstats.com/r/askreddit
But if I actually visit the subreddit, I see many posts from within the last 24 hours with 300+ comments each. Is it possible subredditstats.com had some kind of regression in its counting around that time? It could be related to the API changes.