In the vast ocean of programming and technology, we all strive to stay afloat. DEV has emerged as one of the foremost platforms for developers to s...
For further actions, you may consider blocking this person and/or reporting abuse
We have some big improvements on this front rolling out.
Integrating some much smarter AI tools to explicitly favor less click-baity content and separate signal from noise. Won't be a silver bullet, but is the biggest improvement we've ever made.
Given the nature of the industry, a significant portion of community members is likely composed of beginners.
It seems like "Top X" is a magic keyword that attracts them.
How can we distinguish such posts?
Classifying them through arbitrary judgment might introduce overly subjective opinions into the algorithm.
I am sure the algorithm would neither be arbitrary nor subjective. I think it would be great.
And everyone appreciates quality in the long run, even beginners.
Will it recognise AI written/assisted articles... like this one?!
Generative AI has been proven it canβt recognize the patterns of its own output with any reliability. I would not count on it even if it did.
There are other AI-written/assisted articles out there you need to be bothered about. The author of this post is spilling facts.
Hey RafaΕ! You've offered a really thoughtful breakdown β appreciate you starting up this conversation. π
I just wanted to acknowledge that we are listening in and totally agree that this is an important topic and challenge for us to overcome. We really want to surface the best content we can for folks and agree that the kinds of articles you're calling out should not steal the spotlight.
Anywho, a team member brought up your post in Slack and you can bet that we'll be discussing and thinking about ways that we can improve the mechanisms and procedures we have for uplifting good content to individuals.
I'll keep tuned to the comments here as well as we're very open to feedback in this regard.
You realise an AI wrote most of this article?
AI is just another tool. You probably wouldn't understand some of my English sentences if I didn't use Grammarly. If someone uses AI to generate good posts or even innovative work, let them. But those who abuse AI to create more crap, they are part of the problem.
Nope, I don't realize that. It's genuinely not obvious to me that this was written by a bot.
Perhaps they used a tool to help with their writing but the conversation here feels relevant and about happenings on DEV. I think these are genuinely the author's feelings about the site and I don't think their intentions are to generate articles for personal gain, to promote a business, to deceive others, etc. There is some grey area here and if (big IF, because I still don't see evidence of it) this article was created with assistance from AI, I'd still think its pros outweigh its cons... it's a conversation worth having!
I really don't think it's beneficial to stop this type of discussion from happening or to get too hyperfocused on whether or not the content in the article was created with AI assistance. The primary reasons we worry about folks using AI to write articles are when they're pushing bad information and/or talking about things that they have no idea about, flooding us with lots of low-effort content in a short amount of time, using AI-created content as a vehicle to boost themselves or a business, and/or being purposefully deceptive in an attempt to harm. In this instance, I really am not seeing a downside to this type of post and don't feel like it's worth investigating further because it's a good, helpful conversation for us to have.
I never suggested taking it down, or that it hadn't started a good conversation. I merely suggested that the poster should adhere to the guidelines for such content π
Well addressed.
Thanks for bringing this up. This has been an issue for DEVβand frankly every other aggregator siteβsince time immemorial. The frustrating part about this (evergreen) trend is the fact that it works! Now, I'm not sure which way the implication goes:
Personally, I subscribe to the latter implication. I do not doubt that the DEV team has been combatting this issue for years now. In fact, this is far from the first #discuss post that raises the same issue. As far as I can tell, the DEV team has been (and is still doing) their best to clean up the feedβmay it be through tweaking the parameters of the "Algorithm" or simply by having its army of Trusted Members actively downvoting the overly clickbait-y content.
And thus I conclude that these kinds of articles are just inherently rich with engagementβthrough no fault of the platform! I believe this is because:
In conclusion, don't hate the player; hate the game! Listicles are so rampant because they really work! Now, the naΓ―ve solution is to just ban all posts whose titles start with "Top X", but that's unnecessarily hostile and heavy-handed. So what is the solution then?
I have no silver bullet here, but I think it starts with the Trusted Members. Just as the regular folk upvote and engage with low-quality listicles, Trusted Members are equally culpable for the degradation of the platform's health if we fail to downvote them as aggressively. Now, of course, we can automate this process of detecting and downvoting listicles with artificial intelligence, but those are just implementation details.
My overall point is: we all have a part to play in this. It is so easy to blame the DEV platform and the "Algorithm" for something that's inherently highly engaging (e.g., listicles). Such is the nature of the Internet, unfortunately. Facebook has known this fact for decades already. No amount of manual moderation can scale up to the level of curation that we desire.
The way I see it: the most practical way to combat the problem (without AI) is to bolster the presence of Trusted Members. And of course, this does not mean to arbitrarily add more Trusted Members; that would be ill-advised. It all comes down to usβthe usersβat the end of the day. The DEV feed is a reflection of the posts we engage (and disengage) with. If we insist on a DEV feed that's more quality than quantity, we all have a part to play into making that a reality. There is no magic silver bullet for this problem that requires zero human intervention.
I totally agree, well worded! It's also infuriating for advanced developers, that struggle to find quality articles.
Yup, ChatGPT words things very well sometimes!
It might, but I don't really care if it is. This is something that needs to be addressed in one way or another. If AI convinces the mods and community to do something, then it's a good thing I guess.
I could not agree more. This community needs to stop making stereotype posts and make some good, quality posts like the experts DEVs they are.
I seriously had to set my level up to βExpertβ just to even see something even remotely to a fairly new DEV like me.
The worst thing about posts like those is that they WORK. People look at them, they get popular, and real content doesn't get the attention it deserves.
Thanks for making this post; in my opinion, it very clearly states the issues on this community. As you said,
Remember, in the grand scheme of things, being a great developer is not about knowing every tool or resource out there. It's about problem-solving, creativity, and continuous learning. Let's focus on fostering these essential skills, and the rest will follow.
Best Regards!
Best_codes
This is a much-needed discussion, so thanks for bringing it up - whether through an AI-assisted post or not. π
There have always been a lot of "Top X" articles, but I've also noticed the recent increase. What's worse, I think some are playing the DEV algorithm to their advantage.
Many of these articles have their reactions/bookmarks "maxed out", i.e. it's very visible that a few users gave all 5 reactions possible, and also bookmarked the post. On top of that, these stats grow rapidly in the first hour of the post being published. That's not a typical behavior and likely indicates a group of people trying to prop up the post (e.g. for marketing efforts). This, in turn, puts the post at the top and drives the actual engagement.
Now, asking your friends and colleagues to upvote your post isn't something bad in and of itself, but it seems to be an organized and repeated effort, that seems to feature articles of lesser and lesser quality over time.
On the other hand, I'm also not fully opposed to "Top X" articles. I found them useful many times, whether looking for a tool in a new category or doing some research.
I'm also guilty of writing a few of my own, though, in recent times, I tried to both limit it and, when writing one, focus on a few options only, in greater detail. I feel like such "Top 3s" can deliver more value and insight than "Top 50", which is not much better than a Web search.
All in all, the DEV algorithm should be changed to prevent these articles from getting to the top of the feed so often. Some possible suggestions:
If these changes work, fewer of these posts will be getting to top, there will be less incentive to write these, and there will be a greater variety of authors and topics reaching the top of the feed.
Even though I know that they're clickbait, they always have so many emojis and awesome thumbnails that I can't help but read them. It's so annoying. (I do still read other articles, though.)
Thanks for writing this insightful post @rgolawski! :) FOMO is a major reason why I've considered writing "Top X" style articles, but what has stopped me from doing so is they generate conversations that tend to be superficial (e.g., I think X is a better source for learning JavaScript). So, I wonder what can be an alternative to this type of article?
Our posts should be a documentation of the beautiful journey as a developer.