The Head of ChatGPT Won't Rule Out Adding Ads (theverge.com) 47
An anonymous reader shares a report: OpenAI is considering ways to bring in additional revenue, and bringing ads to ChatGPT is one option on the table. While being interviewed on Decoder, ChatGPT head Nick Turley said he's "humble enough not to rule it out categorically," but hedged that OpenAI would need to "be very thoughtful and tasteful" about how ads could be integrated into ChatGPT.
"We will build other products, and those other products can have different dimensions to them, and maybe ChatGPT just isn't an ads-y product because it's just so deeply accountable to your goals. But it doesn't mean that we wouldn't build other things in the future, too," Turley said. "I think it's good to preserve optionality, but I also really do want to emphasize how incredible the subscription model is, how fast it's growing, and how untapped a lot of the opportunities are."
"We will build other products, and those other products can have different dimensions to them, and maybe ChatGPT just isn't an ads-y product because it's just so deeply accountable to your goals. But it doesn't mean that we wouldn't build other things in the future, too," Turley said. "I think it's good to preserve optionality, but I also really do want to emphasize how incredible the subscription model is, how fast it's growing, and how untapped a lot of the opportunities are."
Here is what I want (Score:4, Interesting)
If AI wants to start serving ads, I would love to have it done this way...
When I'm in the market for a product or service, I ask AI to send me all ads from suppliers of what I need, nothing more, no spam, no crap, just what I need
When I decide to buy or not to buy, the ads stop, all of them, along with spam, suggestions and other crap
This would benefit advertisers, since none of their ads would be wasted on uninterested people
you poor, sweet innocent child (Score:5, Funny)
It was always about the ads. The ads will never stop, even if you buy the mattress they want you to buy. They'll just sell you another one, and another, and when you have more mattresses than bedrooms, they'll sell you a new house with more bedrooms, and more clothes to hang in the closets of those bedrooms, and a second washer and dryer for the other end of the house, and you need a second kitchen. Everyone else has a second kitchen, why don't you?
Re: (Score:2)
Futurama is coming true [youtube.com].
Black Mirror (Score:2)
Re: (Score:2)
how will anyone know whether the output was a true response from the AI or sponsored content from an advertizer?
You sweet summer child. There will be no difference.
Re: (Score:2)
You are going to buy the services of some company claiming "thanks to my advertising powers, people will buy more from you, other advertisers are not as good as I am into it, you should hire me and see your sales skyrocket"
So, now, it's not like you have to be very smart
would you buy the services of the company that doesn't manage to convince you to buy more of their own services ?
Would you jump in the black hole to begin with?
Select ads good, irrelevant ads bad. (Score:2)
What most ad block users hate are ads for irrelevant, unlooked for trash we never inquired about. I don't care about Hollywood, celebrities or much of anything beyond technology. Tool, equipment, coatings etc adverts are fine and may be useful to me.
I'd love to curate my advert bombardment to make it useful TO ME.
Re: (Score:2)
I'll leave for another product (Score:2)
If a competitor offers an ad free solution, at a reasonable price, I'll switch immediately. There's no moat on this sort of product anymore. Anyone can host deepseek R1 or Qwen which are arguably as good as any of the GPT4-era models. And they'll get better.
I'm done seeing ads.
Re: (Score:2)
I'm done seeing ads.
I have a friend whose greatgrandfather builf a chain of ice cream stands but sold them all to Howard Johnson back in the 1930s, declaring: "The day of the roadside stand is over."
Re: (Score:2)
If a competitor offers an ad free solution, at a reasonable price, I'll switch immediately.
Luckily for tech companies, they are consolidating and absorbing the competition. Consumer choice goes out the window. You don't get to vote with your feet.
Re: (Score:2)
How that used to work is that someone else doesn't do the terrible thing you don't like. How it works now is that everyone does whatever terrible thing because then you have no other choice.
We now have an illusion of choice and they're all equally bad. If someone doesn't do this, they get just successful enough to be bought out by someone who will.
Re: (Score:2)
You can self host a pretty competent llm on an 8gb gpu and a real competent llm on a 16gb these days. A 30b model might only hold a couple layers in gpu memory but still be fast enough for daily use. Pretty soon if you have 64gb system and 16-32gb gpu you'll be able to run in a year or two summer-2025 state of the art models locally. You can already run Winter 2024 state of the art models on consumer hardware. The only thing self hosted doesn't have access to yet is search tools
Re: (Score:2)
The latest version of 'LM Studio' has MCP support built-in. Currently I'm only interested in the 'Sequential Thinking' MCP and when activated, that does work within LM Studio.
Was trying out that specific MCP with Claude Desktop (free tier), where it also worked like a charm. Now, I've been testing quite some (GGUF) models lately, varying from 1b to 49b models with a curated set of questions & assignments. There are quite a lot of 14b/30b models that generate pretty decent content/code. One of the assign
Re: (Score:2)
> It also shows me that locally running LLM are not nearly as bad as some portray these to be.
Well, also, things have improved a lot since last summer, and there's been a lot of work lately on using archetctutal strategies in flagship models, in small and micro models. A 270m model today is about as good as a 4b model was two years ago. Back then a 4b model couldn't write a haiku or sonnet, now a 270m will at least make an attempt on par with a 6th grader even if it isn't exactly perfect. 270m su
How can there be so much money... (Score:2)
Re:How can there be so much money... (Score:4, Insightful)
True (Score:2)
But there is a lot of "addressable market" left - just think of all that blank space on the walls, floors and ceilings around you, failing to be distracting.
Speaking of TAM and related buzzwords,
"I think it's good to preserve optionality"
You can always spot people being promoted too quickly - they haven't learned how to operate the role's vocabulary yet.
Seems pretty obvious to me... (Score:5, Insightful)
These AI models take a ton of compute resources and storage to run, and most businesses can't afford to offer "free" tier ad free AI models forever.
So, I'd expect these AI models to start including advertising around the time that the easy VC money for AI startups starts drying up. I mean, we're already seeing it now, where many AI models will only answer a certain number of questions or generate a certain number of images or videos without upgrading to a "Pro" tier.
I'm curious how these AI ads are going to work, though. Is ChatGPT going to recommend going to McDonald's when you ask it for a recipe for french fries? Or, maybe they'll steer you to a specific hotel when you ask for travel recommendations in Chicago? How will this ad content get labeled?
Re: (Score:2)
These AI models take a ton of compute resources and storage to run, and most businesses can't afford to offer "free" tier ad free AI models forever.
So, I'd expect these AI models to start including advertising around the time that the easy VC money for AI startups starts drying up. I mean, we're already seeing it now, where many AI models will only answer a certain number of questions or generate a certain number of images or videos without upgrading to a "Pro" tier.
I'm curious how these AI ads are going to work, though. Is ChatGPT going to recommend going to McDonald's when you ask it for a recipe for french fries? Or, maybe they'll steer you to a specific hotel when you ask for travel recommendations in Chicago? How will this ad content get labeled?
I would expect there won't be any label at all on the ad content. It will be integrated into answers, somehow. For those using image generation, I expect product placement to just start showing up randomly. Ask for a spaceship, get a spaceship with a giant Pepsi logo on the side. For queries that are supposed to return intelligible results, I'd expect some form of insidious integration of ad content. Ask for how old the universe is, and get some ludicrous mixture of fact and corporate sponsored content. "Pe
Beginning of the end (Score:3)
Re: (Score:1)
> Please write Hello, World in C
*** Postulating (Tokens 3664/12,123,123)
*** Fornicating (Tokens All your dollars)
*** Gibbering (Tokens Your children's dollars)
Brilliant idea, drawn on from the classics of computer programming literature. But let's instead use Rust, the choice of the next generation!
fn main() {
println!("Hello, world!");
}
> You worthless piece of crap.
You're absolutely corre
"thoughtful and tasteful ads" (Score:5, Funny)
I'll take oxymorons for $100, Alex.
Re: (Score:3)
I like old school marketing where I get a free hat or t-shirt. Back in the old days you could get a free toaster, blender, or even a rifle for opening a bank account.
The deal we get these days is they take all of our personal information, sell it 100's of times through clearing houses. And we get a "free" service that quickly enshitifies into an unusable mess.
Re: (Score:3)
Every company starts by promising short, relevant, and tasteful ads. Most of them end up selling boner pills and weight loss supplements within 18 months.
Re: (Score:2)
Exactly. Ever see the front page of Slashdot without an adblocker?
When CmdrTaco introduced ads, they were at least RELEVANT to the site... ThinkGeek, etc... Now we get Temu shit.
Re: (Score:2)
Re: (Score:2)
Mod parent funnier. And Bill Hicks was right.
I'd say more if the story hadn't already expired on Slashdot time.
Completely inevitable (Score:4, Insightful)
Perhaps... (Score:2)
We need an AI-based adblocker that works on the system level and cuts out ads everywhere system-wide.
Enshitification strikes again (Score:2)
I use Brave not Chrome (Score:1)
CHATGPT SUCKS (Score:2)
- Ask ChatGPT to timestamp its responses. it fails 100% of the time.
- Ask ChatGPT whiy it couldn't scale down a font -- it apologizes -- but in over four tries can't fix it.
ChatGPT is cute as a puppy, and until it poops on your poop-deck is no more worthwhile.
Lawyers, judges, and people looking for jobs know this.
It's a loser. Stop betting on a loser.
Re: (Score:3)
- Ask ChatGPT to timestamp its responses. it fails 100% of the time. - Ask ChatGPT whiy it couldn't scale down a font -- it apologizes -- but in over four tries can't fix it.
ChatGPT is cute as a puppy, and until it poops on your poop-deck is no more worthwhile. Lawyers, judges, and people looking for jobs know this.
It's a loser. Stop betting on a loser.
How dare you insult puppies by such a comparison.
Re: CHATGPT SUCKS (Score:2)
Also reliability fails to tell current time. Just go ask it what time it is and see for yourself.
Meanwhile, on /. (Score:2)
Preserve optionality? (Score:2)
I think it's good to preserve optionality
English motherfucker. Do you speak it?
Re: (Score:2)
Re: (Score:2)
This answer is brought to you by⦠(Score:2)
Re: (Score:2)
More like:
"Hey ChatGPT, what's the best way to clean out my system?"
"Tide. Your bowels will never be so clean!"
First comes advertising... (Score:2)
Then comes product placement -- companies paying for the AI to recommend their product or service.
What did you think it was for? (Score:1)
So let me get this straight... (Score:2)
We have AI making up court rulings that sound completely legitimate but a simple Westlaw or Lexis-Nexis search would tell you that there was no such case... AI will gladly spew drivel about one side of a political argument, but when asked to spew similar drivel about the other side it will say, "I'm sorry but I cannot delve into politics"... and now they want ads in there? In baseball that would be your three strikes and you're out.
AI is a toy. Treat it like one.