Hacker News Reader

Reddit has possibly the worst mobile experience I've ever seen. I've never seen a company try to force its (awful, slow, buggy) app down people's throats harder if you try to use the mobile web version.

The entire mod culture needs to be reworked too. They need mods because more people brings more trolls, but aggressive banning and simultaneous muting so people can't ask why just leads to more trolling. Not to mention most mods just cite every ban in the coverall "read the rules / figure it out" so they have to spend no effort, even though those rules don't cover the actual ban reason.

I just can't believe how many consistently bad decisions are made there. The biggest of all was probably removing the separated and accurate upvote / downvote counters which is a massive feature to show the strength of otherwise downvoted comments.

Aside from these feature complaints, how does a company of this size and importance have such an astonishingly badly coded interface? I always have to use the Old site. It reminds me of an intern coming in and on day 1 marking all code as "legacy" and saying "Electron is the greatest thing ever made, because it's easy and just works. There's literally no reason to use anything except Electron for any project."

Removing downvotes is largely something advertisers demand. Advertising is still stuck in some old ways. I know some in corporate ad departments that feel the need to respond to every negative comment on the internet. That is implementing hell for anybody that has that task.

> I've never seen a company try to force its (awful, slow, buggy) app down people's throats harder if you try to use the mobile web version.

Instagram and Twitter do this as well, FWIW.

Instagram and Twitter's apps are a lot better than Reddit's.

Taking shots at the engineers is too far. Reddit is about a million times more stable than it used to be a few years ago. Also I don’t think Reddit or the Internet needs “saving”. They’re both more successful than ever by any common metric whether we like the direction they’re going or not.

How do you quantify that it’s more stable?

Backend sure, it’s more stable.

What’s on the front is inexcusable though. This is not just dark patterns: shit straight up doesn’t work. Good engineers don’t ship that.

The video player has sucked for years so much that I built a service that fetches the raw hls URL and plays that directly in the browser. Night and day. How do you mess up something that works out of the box?

My experience only having used old Reddit is that it’s stupidly slow all the time and it goes down at least once a week for me (I’m not looking at it all the time…)

I can’t think of a single website with that many backend problems.

I have a 10-year-old reddit account. When it was static pages served out of lisp, it was rock-solid. Their rewrite got to the instability you're describing, and it never really recovered.

The Twitter account named Reddit Status was started in 2010 to track Reddit’s downtime. It has always been unstable. I’ve always thought it was strange how often Reddit was crashing or going down for scheduled maintenance when no other websites ever seemed to do so.

They moved to python 16 years ago

The Python rewrite happened 16 years ago. 10 years ago, there was probably an order of magnitude (or more!) less traffic and this is probably what causes most of their instability.

10 years ago was about the time when they first moved to Cassandra and that was a pretty rough period. Lots of rose tinted glasses around here. Reddit has had plenty of stability and scaling problems during the early years.

Kinda: http://www.aaronsw.com/weblog/rewritingreddit

I'm going on 2 months now, where I cannot login to my account on 3rd party android apps. I can using an alt account. I'm not banned, or anything. It's just a known problem -- probably dealing w/ a cookie, or setting, or variable...

I've tried turning on/off 2fa, resetting my password, etc... nothing works. i've contacted reddit and submitted to their help sub, crickets.

So yeah, taking shots at the engineers imho is fair game, this is a major issue. I can't even go to /user/prefs/apps to enable/disable apps attached to my account which seems like it could be a security vulnerability. I always get "You broke reddit"..

when I try logging in on Relay or ANY other 3rd party mobile app it hangs after I'm logged in and can't handle the token exchange part of the oauth flow where the data is passed back to the app.

If you call this stable.... I'd hate to see unstable. The mobile app is a no go... I like to be able to change views between small cards, gallery, and large-card depending on what I'm using reddit for...

It sounds like you're mad at explicit decisions made by product owners, like the content on the frontpage or dark patterns that push mobile app usage. I guarantee none of the rank and file engineers are responsible for those decisions, and they almost certainly detest them just as much as you. The problem is if they don't implement them then they don't get a paycheck and can't put food on the table. It's silly to paint them with blame.

They’re all Bay Area engineers. So switching jobs is an inconvenience with a raise on the other side. The changes aren’t their ideas but it’s not like they can’t refuse. They can just work anywhere else. They probably don’t care.

I mean… browse through HN for the day. According to its users you shouldn’t work at any company that pays well because they are all evil. At a certain point you just don’t care what people are writing on forums.

If HN had its way, we'd all program our servers using common lisp using emacs, serving pages with zero CSS or javascript.

There's a bunch of them that don't seem too bad: netflix, spotify, discord. Maybe even apple, depending on your definition of evil.

Spotify? Really?

Yes its truly awful for you to be held accountable for your place of work and its impact on the world. Completely unacceptable.

I hate making a ton of money and giving everyone in my family a better life. Really sucks.

The things you are holding people accountable for are very questionable in almost all cases. In most cases it just comes off as jealousy that someone is doing better than you.

I’m all for being held accountable. But i put very little weight in public opinion being something that we should even think about at this point.

> In most cases it just comes off as jealousy that someone is doing better than you.

That is certainly a very convenient way of dismissing something you find uncomfortable.

Your last argument can be applied to about anything regarding performing unethical tasks to stay afloat. I don’t find it convincing for engineers that work for Reddit. They can start work at any other tech company, too, so by absolving all the blame is a bit much, in the end they choose to implement it right?

Well, at least they're not working for Palantir or Meta, so give them some credit.

Work at any high profile internet company and people on the internet will complain about something you’ve worked on.

There doesn’t seem to be anything particularly unethical about any work Reddit engineers have done.

I mean, sounds like you have a problem with the product managers not the engineers.

> This is what happens when you hire mediocre engineers

Nope. It's the result of perverse incentives and poor management. I can imagine the developers (even the designers) had a good vision of what to build and how to build it, only to be overruled by management with ideas like "make it janky as hell so we can force all our users to our mobile app to track them better".

I hope that no single organization replaces Reddit. Let a billion independent forums bloom. For the sake of discoverability, it would be nice if a decent subset of them implemented more or less the same API that search engines could call from the outside and that their UIs could expose to users so that users could easily move from forum to forum. But discoverability is double-edged - it builds size but it also brings Eternal September and homogeneity.

As for structure, I think that forums and image boards that are basically just chronologically ordered items of text where each item of text can have either implicit or explicit links to previous items of text are in a sense superior to tree-based forums like Reddit anyway because directed acyclic graphs are a superset of trees.

As for upvote/downvote mechanisms, I have never seen a forum that was improved by them. Such mechanisms are probably good for driving engagement but from the point of view of stimulating thought, I think that as opposed to having an outside algorithm preference certain comments over others, it is better when users have to either manually go through a flat chronologically ordered list of comments or implement their own search functionality on top of that list in order to find comments that they want to engage with. If a forum provides advanced tools to search and sort comments, that is good, but I would prefer that a forum not build any given sorting other than chronological into its default UI.

Reddit, at least, still allows users to sort in ways other than by "best". In my experience and according to my taste, large subreddits usually are often pretty useless unless I sort by "new" or "controversial". Sorting by "best" all too often - not always, but all too often - just brings the least common denominator up to the top. Another good method for using Reddit is to make note of any individual users whose writings you find interesting and to then just read through their Reddit histories - this can also be a good way to find interesting subreddits that you might not have come upon otherwise.

Interesting idea. Tapatalk is the only thing I've seen that attempts to unify UX for different forums

> Let a billion independent forums bloom.

They did. Your first paragraph describes Reddit perfectly.

> it is better when users have to either manually go through a flat chronologically ordered list of comments or implement their own search functionality on top of that list in order to find comments that they want to engage with

Which website(s) are you thinking of?

> now they're trying to get into everything, and can barely execute

I feel like this happens to a lot of organizations/products that start off being really neat.

Fast forward 10 years, and most/all of the people responsible for the good stuff have moved on. The brand ambles on like a relative with dementia. The lights are on, but you know nobody's home.

Amazon is the only one I know that didn’t.

Honestly, I think their webstore-side is surprisingly weak. As experience that is. Ofc, logistics and so on and AWS is still moving.

Amazon the web store used to be good. It had good search and it was nigh impossible to get something counterfeit from it. Nowadays it's all people talk about.

I think Amazon is very much a victim of its own success and rapid growth.

They built and presently operate a data center at Langley for the CIA.

They started as a bookstore.

They started out as a bookstore but Bezos had experience in finance, they quickly sold media, and the warehouses were nothing like other bookstores and it looks like it was more a base for larger operations that didn’t anticipate staying a bookstore, the same way their online functionality didn’t stop at a website, AWS and selling other stuff came fast.

Sure they were a bookstore first but they over engineered it to scale.

Ah the myth of the mediocre engineer. It’s always the other team at the other company, too. Those engineers sit in their open floor plans, designing reddit?

(highly opinionated)

I always find it useful to compare reddit to imgur. IMO the future of reddit is the present of imgur. The value proposition of imgur in large parts is the same as reddit IF we ignore the part where imgur was hit with a deadly blow when reddit launched in-platform media hosting.

Imgur's push for mobile app and the desparate attempt (I will say plea) to form a loyal community around the product really shows how internet platforms fails to innovate and think community is the only to tool growth and survival.

On the other hand we see another example of Tumblr where communities get in the way of operating a business.

After everything I say I find Facebook and YouTube is successful because community driven platforms work in niches but to scale up you must establish yourself as a monopoly and focus on running a business while juggling all the negativity.

I remember when Imgur was launched as a free image hoster for Reddit users. It was jarring to me, years later, to see it had attempted to become some weird Reddit/9gag mash-up. Not because it was a bad move, business-wise, but because it became clear how unmonetizable the plan was to start with.

That said, I agree with you: it feels like Imgur makes the Reddit move about twelve months before it, every time. From gif hosting to discovery to live feeds to requiring logins for NSFW content, Imgur basically leads Reddit. It honestly makes me wonder why I bother with Reddit at all...

> So the trillion dollar question is, what's next to replace reddit

Discord is taking some of their traffic. But it sucks as a replacement since the threading model doesn't match a true forum.

Original creators of Reddit were great but now they seem to have shitty devs/management. Interviewed with them and wasn't impressed.

I have clone but it needs little more polish and timing :)

Nice to see you on HN. ;)

It used to be written in lisp. It used to be good before eternal September. Being a redditor was a nerdy virtue signaler. Now? You can be seen as a white supremacist, a commie, a clout chasing loser trying to get upvotes, or an incel when you mention it.

I’ve been reading Reddit for fifteen years. (A sentence that should never be written.) Their engineering has always been terrible. Even when they were much smaller than they are now, their site constantly crashed. It’s remarkable how little improvement they’ve made in over a decade. There are quirks, like how you can only see a maximum of 1,000 results, whether on subreddits, searches, or user profiles. This limit is apparently unalterable because of legacy decisions.