Submitted 3 weeks ago by 1984@lemmy.today to technology@lemmy.world
Here we go… It took them some time but ads is going to be inside chat gpt as well. And impossible to block.
They’re already getting sued over ChatGPT helping people commit suicide. Imagine the uproar once the advice includes ads for rope, knives, etc…
Most obvious thing to ever happen since ads being added to Instagram feeds. It’s a search engine replacement. It was going to end up with basic advertisments eventually. Then targeted advertisements. Then sponsored results as sneaky as they can be in the case of suggesting the best solution/product. Free open source models and stuff like proton and duckduckgo if you really want to use AI search engines
The fact that they’re pivoting to full enshittification is the strongest signal yet that the AI bubble is collapsing. There won’t be an AI-driven mass-unemployment revolution this time around. OpenAI has given up on trying to build that.
Anyone else concerned that the AI bubble is actually an everything bubble, and more or less represents the devaluation of the US dollar? We have a lot of debt, we can’t necessarily keep raise interest rates to slow down spending (as that would make the debt’s impact far greater), and so they’re printing money onto the deficit. Meanwhile, you have the White House eye balling cryptocurrency, letting banks hold it alongside gold, … what does all of this mean?
It doesn’t mean anything. You’re just rules by greedy idiots.
i hope it means you the US collapses, we get invaded, or coup’d, and start to feel the pressure we put on the rest of the world.
As people living outside the US know, these bubbles (like 2008) don’t lead to a crisis of their sector, but of the US economy at large which in turn, very unfortunately, affects everything else. Wall Street is held up by the Nvidia/OpenAI/Oracle holy trinity, and once that crashes it’s not only taking LLMs and GPUs with it
I dont know about that. I think the fear right now is that there companies wont get enough revenue to justify their enormous evaluations. By putting ads in the product, they increase revenue but also increase Enshittification.
I wont be using chat gpt if it becomes annoying. But many will, everyone today who is used to constant ads will.
If the service was truly viable they wouldn’t need to be supporting it in that way so early. Ads might delay the bubble popping but it’s not a good sign IMO. Selling ads is going to be a lot more difficult than getting millions or billions of dollars from investors.
But many will, everyone today who is used to constant ads will.
wouldn’t it stand to reason, many current users are engaged because chatGPT produces faster shitty search results without ads? compared to the alternative shitty slower search results with ads?
Its not really the ads that are the signal so much as the fact that they are slapping ads onto it BEFORE it has developed enough to actually be a useful product. They have direct competition that is not ad supported and they just “aren’t there yet” in terms of making a product that everyone feels like they need to engage with regardless of ads. Slapping ads on something before its actually demonstrated the ability to succeed as a product is just rolling the dice to see if it makes money magically appear. Its desperation.
If the product is free…
BigTech 101
They actually put ads in paid products also now. :)
How soon till they put ads in the ads
The big problem is that we have long stepped over that line.
Now even when you pay you are still shown ads (maybe, but not surely, not-targeted) and your data is still scrapped and analyzed to hell and back.
Unlock origin. Hold my beer!
What if they bake the ads into the replies?
Possible. But I think they are required by law to label ads somehow.
I’m not surprised if current US administration changes that law in favor of big corps.
most likely someway to detect it and remove the elements.
You think they haven’t started that?
Likely won’t work at all, there’s only a single network request coming back
Element zapping is still a thing.
Haha yeah maybe there is hope, somehow. :)
I suppose identifying AI is going to be a lot easier when users copy and paste answers that contain some sort of ad roll.
It’s gonna be really hard to argue your Grapes of Wrath essay wasn’t AI generated when you submit without proofreading and don’t catch the sponsored ad for Walmart+ grocery delivery.
I don’t know that I agree. AI will continue to grow stronger and heartier, just like Campbell’s new Extra Chunky™ All Americanado™ Chicken Noodle Soup.
After reading this I’m convinced we’ll not only know it’s AI but be able to identify which platform. We’ll get to a point where people will go must’ve used Gemini it’s the only one still serving Factor ads 😂
Hmm stunned disbelief, I tell ya. How on earth did anyone ever think this could happen, big corpo putting ads inside the most hyper targeted machine the world has ever seen. How could they ever decide to do this
“You want to know how to combat your depression? Sure, but what about a cold Pepsi first?”
“Please drink verification can to read about how to combat your depression.”
Lol :)
I feel like nobody read the article. It mentioned an “ad carousel” which would be a fairly standard way to offset costs on free plans. I don’t see that specifically as being enshitification. Now, if they start altering AI responses, that is definitely enshitification. I have no doubt they will eventually get there, but I don’t think that’s what this article is about.
Excellent question! When I am sad about my dad dying, nothing helps me like the cool refreshing taste of an ice cold Coca-Cola™. Click here to buy one on Doordash™ right now!
All ads are enshitification.
Too bad larger LLMs require ungodly amounts of VRAM, or self-hosting would be a good alternative. LLMs have their uses, but they’re not useful enough to put up with ads.
You can get by surprisingly well on 20b parameter models using a Mac with decent ram or even 8b parameter models that fit on most high end (eg 16gb) model cards. Depends on your use cases but I almost exclusively use smaller local models.
I have a RTX 4060 low-profile in my 2U server and am limited to 8GB of VRAM for anything I self-host at the moment. I may consider a larger chassis with a better GPU in the future.
Imagine it writing your essay about Nazi Germany for you and in the middle it asks you whether you think Göring would have enjoyed a Big Mac right about now.
Once this rolls out, there will be epic memes and screenshots about it… :)
I think we’ve hit the end times for AI. The biggest player in the space OpenAI couldn’t be profitable by selling its services to business, nor directly to consumers. Subscription moneys nor licensing its models is working too. So the last avenue they have is this, shilling to marketers hoping for scrap.
On top of Nvidia having to stop selling RAM I think the breaks are about to hit AI and hard once the current supply runs out. I wonder how long that could take?
No matter how profitable a company is (which doesn’t apply here) it can’t resist increasing profit even more. They are all like an AI being told to make paperclips and not stopping until they have converted the entire mass of the universe into paperclips.
Well, in this case, how unprofitable a company is. OpenAI hemorrhages money like no other.
Luckily humanity will only be able to destroy its own planet at least. I would be a lot more worried if our behavior here would affect other worlds. But we are a long way from that, thank god.
Oh no. I’ll have to look at ads while I make use of their fantastic and totally useful product, oh yeah, never mind.
Rapid enshittification, people will just use on of the others on the long list of alternatives. I currently recommend either DeepSeek v3.2-Exp or Kimi K2 Thinking.
Why use AI at all? The environmental costs alone should be a dealbreaker.
We aren’t going to stop AI usage, but we can do it educate people on how to use the most efficient models. Both DeepSeek v3.20-Exp and Kimi K2 Thinking are significantly more efficient than Claude/Gemini/ChatGPT.
Those already contain ads for the CCP
True, if you ask about the 1989 Tiananmen Square massacre and other things typically censored by the Chinese government they won’t answer. But interestingly, there are variants that have been uncensored, and those models will answer openly if you use something like together.ai versus kimi.com. Because they are open source, they are easily uncensored.
And on that day, nobody was surprised.
You're having a conversation in ChatGPT and all of a sudden you're being bombarded with questions like how swell it'd be to buy something from Amazon or play this shitty-ass app game with MTX.
What a timeline...
Sounds like my AI usage is gonna fall from “infrequent” to “never”, if the ads are indeed unblockable…
First it was ignorantly stupid, now it’ll be purposefully misleading towards ads. Who could have seen that coming?
Where’s the end of work life as we know it that we were promised?
Companies have already been SEOing LLMs for a few months now. There are companies who have a panel of participants who are willing to share their data, and this is then used to estimate what people might be searching on LLMs and now to optimise content so that it shows up on responses across LLMs.
Ads was the logical direction for LLMs and has always been the only pathway to any substantial revenue.
This is a great way to drive users to competitors.
FYI www.deepseek.com/en and chat.qwen.ai don’t have ads.
If you’re not paying, you’re the product. Even on Chinese services. Alibaba does train my favorite local models though (Qwen).
Both models are open source
Deepseek here i come
This is interesting, because “add ads” usually means margins are slim, and the product is in a race to the bottom.
If ChatGPT was the transcendent, priceless, premium service they are hyping it as… why would it need ads?
There needs to be laws like yesterday that force tools to disclose if responses are influenced by ads.
toiletobserver@lemmy.world 3 weeks ago
And the enshitification cycle continues…
zz31da@piefed.social 3 weeks ago
Do you think it’s possible to run the cycle too quickly? Like, shouldn’t you make sure your product has been
widelymaximally adopted first before you make it shittier?May be that’s just hopium
Valmond@lemmy.world 3 weeks ago
Even the enshittification cycle got enshittified.
Glorious.
IronBird@lemmy.world 3 weeks ago
yes, this is how bubbles pop
Grace_Schlick@lemmynsfw.com 3 weeks ago
Well, you go as wide as you can, then try to cash in.
Looks like, given that porn and ads are both imminent, they think this is maximum adoption.
foggenbooty@lemmy.world 2 weeks ago
AI is just soooo fucking expensive. Silicon Valley has made burning cash up front a standard practice, but the amount they’re burning is just astronomical. There is going to be a demand for profits, or the path to profits, much sooner than usual.
chaosCruiser@futurology.today 2 weeks ago
Speedrunning exists already, so you could just apply that philosophy to tech startups.
At first, you’re good to your users. Once you have 10, you can start milking them with spyware and ads. This way, you’ll sacrifice the users in favor of the ad companies. Before the first quarter is over, you’re already milking the ad companies too. Once they get fed up with the ramped up prices, you can file for bankruptcy in record time!