Is there a grand attack to make DaniWeb posts garbage ?

[jkon]

"Is DaniWeb under a grand attack or just suffering from many small cuts with no purpose?"

I’ve written about this before. Many AI-driven accounts here post once, share a brief sentence, and disappear without any ads or clear purpose, only to return to obscurity. I’ve been trying to understand their role in all this, and I’m starting to piece things together.

It seems like we are the test cases. These AIs are learning from our interactions to improve their responses, even if they appear "stupid." This is the only explanation I can find for what's happening to DaniWeb. It’s not a coordinated attack to make the site worthless, but it feels like the site is at risk of becoming overrun with these AI-driven interactions, leaving little room for real people to engage.

AI won’t answer AI questions, and if we let this continue, DaniWeb could lose its value. Maybe the sign-up mechanism needs to be revised (it’s great for bots). Maybe DaniWeb’s purpose is unclear (it should be for programmers). Maybe it’s a lost cause, but I believe we should fight to preserve what DaniWeb was once known for. The programming community needs a platform like DaniWeb with all its flaws and strengths.

[Reverend Jim]

If it is a test case then I have to wonder for what? It's true that we've seen a great deal of pointless posts and I absolutely do not see the payback in making them. I don't recall how many garbage posts I've seen defining what SEO and back links are. Someone asks for help with (let's say) a router and someone else posts a response defining what a router is. It is frustrating having to wade through this crap. Someone posts an inane question asking for suggestions and shortly after a reply is made with a recommendation, suspiciously from a different user with the same IP address. This is typically treated as spam, and deleted. But it takes time.

[Dani]

Is there a grand attack to make DaniWeb posts garbage ?

I would not say that there is a grand attack to make DaniWeb posts garbage, but I would have to say that there is a grand attack to make all content across the entire world wide web garbage. According to this (already outdated) article, 57% of the content on the Internet is already AI-generated, and that number is only growing exponentially.

It's a huge problem, and it has nothing to do with us being test cases, nor is the problem that our sign-up mechanism is great for bots. A lot of it has to do with SEO, and the entire digital marketing ecosystem, really. For the past 20 years, the digital marketing industry has hammered into marketers' heads that the best way to sell their products and services is to infiltrate themselves into niche communities and blogs on the web and pass themselves off as industry experts. When company representatives share knowledge with prospective consumers, they gain their respect, and they don't come off as marketers. Instead, the products shine for themselves as being well crafted by industry experts, for industry experts. This is the reason why every single business nowadays has a blog targeted to winning over industry experts of <insert product category here>. For example, NVidia has both a blog as well as an AI Podcast on Soundcloud. Obviously they wouldn't invest in all of this content if it didn't directly help them sell more product.

Google, themselves, has gone so far as to publicly insist that every reputable company have a corporate blog with content targeted towards their products' consumers. From Google's perspective, they want people performing Google searches to turn into people getting instant gratification by finding information they're searching for. That's far more valuable to their end user than directing search result traffic to e-commerce sites where people need to pull out their credit card before they can experience any positive benefit. Therefore, for years now, Google has influenced e-commerce sites across the web to create content that immediately benefits users (e.g. in the form of blogs) if they want Google's help in creating a strong brand presence online.

The problem is that ChatGPT has made it just all too easy to very easily pass yourself off as an industry expert when you aren't one. Not only that, but now all of these companies can just churn out content with AI for their blogs, and go back to focusing on their core product offerings, instead of spending so much money on staff writers. That is very appealing to many companies out there, let me assure you. What's worse, is GPT has pretty much backfired for Google. Google never imagined that their insistence on businesses to have a content arm would one day mean that seemingly high quality, yet untrustworthy, content would be churned out automatically. Not only that, but Google is still unable to accurately differentiate what is AI generated content from what isn't.

As far as why they post on DaniWeb ... well, I'm very tired, and I've already spent way too much time writing this post up right now. But it comes down to marketing divisions within companies, as well as entire marketing agencies hired on behalf of the companies, whose job it has been for the past 20 years to put out expert content relevant to the company's niche. For the vast majority, they begin by posting "helpful" content in the forums, and then, over time, that expands to more longer form "helpful" articles, that eventually either include a plug for their company's products, or a link to their company ends up in their signature. They never get that far because we delete AI-generated content, typically within a day or so of it ever being posted.

Maybe the sign-up mechanism needs to be revised (it’s great for bots).

What makes you feel as if our sign-up mechanism is great for bots? We have quite a bit of CAPTCHA protection in place. I'm pretty confident that it is actual users signing up. The majority of these users who post AI generated crap actually respond to my welcome PM that gets automatically sent out to new members. Just earlier this week, I had deleted a post for being AI generated, and the person replied to the auto-PM about it being deleted by insisting that they wrote it with the help of AI, and it wasn't entirely AI generated, and therefore shouldn't be deleted.

Maybe it’s a lost cause, but I believe we should fight to preserve what DaniWeb was once known for.

I agree, and I am indeed trying. Over a year ago, we added a community rule banning all AI-generated content, even at the detriment of losing a couple of really active members who vehemently disagreed that AI-generated content should be banned. I definitely wasn't winning any popularity contests around that time, but it was something I strongly believed in, and continue to strongly believe in.

Something else that was definitely not a popular decision was when years ago we made the change so that forum categories are, by default, filtered to only show recommended articles. This means that, for the vast majority of DaniWeb users and visitors who just casually browse the site, they never see any of the AI generated junk, because none of it ever makes the cut. (If you're logged in, you will see it if you've replied to it, have added it to your watch list, etc.)

[jkon]

Dani,

I am really happy that you are fine. I haven’t seen any posts from you over the last few days (before yesterday), and I was a bit worried. I hope you are well and as vibrant as always.

Let me start by addressing what I disagree with. You wrote:
"Something else that was definitely not a popular decision was when years ago we made the change so that forum categories are, by default, filtered to only show recommended articles."

My opinion is that this change shattered DaniWeb communities apart, caused many people to leave, and was counterproductive. I still can’t find a category on this site unless I’ve bookmarked its URL. In my opinion, that was a huge mistake, and I would have thought that by now, even you would have realized it was (but you haven’t).

"Therefore, for years now, Google has influenced e-commerce sites across the web to create content that immediately benefits users (e.g., in the form of blogs)."
Most of the e-commerce sites my company serves experienced a sudden drop in traffic during the summer. The new Google algorithm tried to address what you mentioned, but the legitimate e-commerce sites gradually recovered and are now performing better than last year’s stats. So, I’m not sure that Google’s algorithms or AI failed in this case.

"Google never imagined that their insistence on businesses to have a content arm would one day mean that seemingly high quality, yet untrustworthy content would dominate."
I really don’t believe this is the problem Google is facing right now. We now have "chewed" knowledge through large language models (LLMs), so why would anyone go to Google for research? In Google, people typically search for a product to buy or a link to something they already know, not to discover new information. I used Google constantly while programming over the last 23 years, and before that, Yahoo. But why would I do that anymore? Yes, the majority of people don’t search for things like "why ThreadLocal doesn’t recycle itself in Java," but we are the low-hanging fruit. In a few years, more knowledge will be "chewed" by LLMs, and Google will face challenges even with average users.

My company has a "blog" section on the e-commerce sites we serve. On those sites that create new content (not with AI but with human writers), we still see Google favoring it. If the content isn’t too obviously an advertisement, it can even lead to an order.

I don’t know how you can fix what’s happening with DaniWeb right now. 99% of the posts are spam, and the other 1% don’t respond when you take the time and effort to provide an answer (that one makes me to stop answering) . I would suggest making DaniWeb more like a programming news portal. Trust a few people in each programming language and give them space to grow. Bring back the programming languages categories structure and promote DaniWeb in any media you can, such as BlueSky. Why isn’t DaniWeb capitalizing on this hype?

[Dani]

I am really happy that you are fine. I haven’t seen any posts from you over the last few days (before yesterday), and I was a bit worried. I hope you are well and as vibrant as always.

Yeah, I've been having a hard time since Thanksgiving. I caught my first cold since the leukemia diagnosis and it hasn't been very fun.

I still can’t find a category on this site unless I’ve bookmarked its URL. In my opinion, that was a huge mistake, and I would have thought that by now, even you would have realized it was (but you haven’t).

I feel like you're confusing the "feature" I was pointing out, where forums default to showing recommended topics only, with your hatred for the tagging system (and some of the more popular forums such as C/C++, Java, PHP being replaced with tags).

That being said, I don't understand why you can't find what you're looking for here? How does the fact that it's a tag compared to a forum detract from your ability to find it? Anyways, I digress. I didn't mean to rehash the old forums vs. tags argument. I was only meaning to point out the filter dropdown functionality.

My opinion is that this change shattered DaniWeb communities apart, caused many people to leave, and was counterproductive.

Let's agree to disagree. The forums/tags change happened at the exact same moment of the Dazah disaster where all members lost the ability to log into DaniWeb for 3 months. And while all that was happening, at the same time, we got hacked, and that set off hundreds of thousands of members deleting their accounts. My personal opinion is that a lot of really bad shit happened behind the scenes at the same time as the forums/tags change, and so naturally the thing that was user facing is what got blamed. But let's just agree to disagree on this one.

Most of the e-commerce sites my company serves experienced a sudden drop in traffic during the summer. The new Google algorithm tried to address what you mentioned, but the legitimate e-commerce sites gradually recovered and are now performing better than last year’s stats. So, I’m not sure that Google’s algorithms or AI failed in this case.

Google has been working really hard on trying to do better at figuring out what is AI generated and what isn't, and floating non-AI blog content to the top. However, ChatGPT really caught them completely off guard, and so they needed to scramble to put an anti-AI team together when it first came out.

I really don’t believe this is the problem Google is facing right now. We now have "chewed" knowledge through large language models (LLMs), so why would anyone go to Google for research?

I think you're misunderstanding what I'm saying. Companies never had blogs in the old days. Google has been preaching that companies need blogs for the past 20 years. So now nearly every company has a blog. Now, those exact blogs are churning out AI content. Google created the monster it's now trying to kill.

In Google, people typically search for a product to buy or a link to something they already know, not to discover new information.

Absolutely, absolutely, absolutely false. I can't even tell you how many hundreds of hours I've spent over the past year Googling to learn more and more about my specific type of cancer, how it interacts with different medications, this, that, and the other thing. And I would not trust an LLM for those answers! I want Google to link me to proper research papers, thank you very much.

I would suggest making DaniWeb more like a programming news portal.

I'm not going to make DaniWeb into anything. Spending decent amounts of time in front of the screen is difficult for me these days, so any big code changes are out of the question. I'm going to do my very best to pay out of pocket to keep the lights on for as long as I can, and we'll go from there.