ChatGPT competitor comes out swinging with Super Bowl ad mocking AI product pitches.
See full article...
See full article...
Should AI chatbots have ads? Anthropic says no.
- Thread starter JournalBot
- Start date
Upvote
151
(162
/
-11)
I would be willing to bet they go back on their word or find a loophole at some point.
What they say and what they do rarely line up.
Also;
What they say and what they do rarely line up.
Also;
Isn't that literally what SEO or AIEO(bingo was his name) does already? So they are influenced by third party product placement they just don't currently profit from it.
Upvote
59
(64
/
-5)
Yeah ok... Anthropic who I'm sure like all the others is acting as a mass survilience platoform is "ad free" for now...
Meanwhile they make you pay for more usage on the "pro plan" unless you wait a week for the usage limit to reset after sending like 15 prompts so whatever.
Meanwhile they make you pay for more usage on the "pro plan" unless you wait a week for the usage limit to reset after sending like 15 prompts so whatever.
Upvote
-19
(6
/
-25)
"Should [fill in blank] have ads?"
No.
Airline flights now have 10 minutes of credit card ads over the loudspeaker. Make it stop!
No.
Airline flights now have 10 minutes of credit card ads over the loudspeaker. Make it stop!
Upvote
118
(118
/
0)
Mostly Ignorant
Ars Centurion
This will be an ad we look back at in 5 years and laugh at, either because AI died in the crib or because Anthropic will have ads. Either way, if there is a there, ads will follow.
Upvote
79
(87
/
-8)
This is like asking if burglars should wash their hands. Yes, I'm very glad you washed your hands before taking my things... now could you please give it back and leave?
Upvote
31
(35
/
-4)
You are absolutely correct, and that is a very insightful observation. You must be fending off night starvation with Horlick's Malted Milk, like anyone of your natural intelligence would be.
Upvote
38
(38
/
0)
All else aside, that's a hilarious ad between competing companies. I suspect quite a few ChatGPT customers will see it on Sunday. Wonder what sort of impact it will have beyond just being funny.
Upvote
5
(9
/
-4)
Upvote
11
(12
/
-1)
Based on their published Claude constitution, I'm inclined to give them the benefit of the doubt/take them at their word on this.
Would be nice to see some reference to that constitution in this article, as it drives a lot of what the company will do on the Claude roadmap.
Would be nice to see some reference to that constitution in this article, as it drives a lot of what the company will do on the Claude roadmap.
Upvote
7
(19
/
-12)
for.the.republic
Smack-Fu Master, in training
I started using Claude and Le Chat last month and they have been great. A lot less artificial chumminess.
Upvote
2
(9
/
-7)
I guess when you're a trillion dollars in the hole, a little ad revenue won't hurt before your bubble pops
Upvote
31
(31
/
0)
It sounds like Anthropic hasn't figured out how to implement ads in a way that will return a profit. So they are letting OpenAI spend money on that research.
Upvote
7
(18
/
-11)
Let's hear what he says when all the investors and VCs who have pumped billions into Anthropic demand their due...
Upvote
22
(24
/
-2)
Won't hurt THEM at least... the amount of power used to generate and insert them will hurt others though.
Upvote
12
(12
/
0)
Ads in chatbots sound so bad.
Like consider how much trouble LLMs have maintaining coherency even now, it's going to be so much worse if you start adding prompts telling them to push certain products.
Remember when we had that article about Grok randomly bringing up White Genocide constantly even in unrelated topics? Imagine that but it's fucking Aldis or something.
Like consider how much trouble LLMs have maintaining coherency even now, it's going to be so much worse if you start adding prompts telling them to push certain products.
Remember when we had that article about Grok randomly bringing up White Genocide constantly even in unrelated topics? Imagine that but it's fucking Aldis or something.
Upvote
29
(29
/
0)
They are shooting for the professional market.
OpenAI's soon-to-be-fatal mistake is that they are trying to shoot for the mass market rather than focus on a niche. Mass market offerings tend to attract the most VC dollars. Nobody cares about a "boring" SaaS offering to pros.
The problem is that there is not much of a use case for non-professional users to pay for ChatGPT, and professionals often have niche focused LLMs that are tailored to their needs.
I'm calling it now: when the AI bubble finally pops, Open AI will be the one that dies. Anthropic, etc, will probably kill the free tier and ramp subscription prices. Google will backpedal on AI and make it more professionally focused.
Upvote
47
(47
/
0)
Personally I think anyone who isn't already making tons of money from other avenues will go under. So google, meta, x, Microsoft will stay afloat just because they can afford it. I do agree that free plans will go and prices will hike. I can see that killing the vibe coding market as well having a further knock on on the smaller companies like cursor. We will also get to see how many of these AI companies are actually just wrappers of other LLMs.
Given how many companies Anthropic are already selling to they must be reaching saturation whilst apparently not profitable. Unless they can survive the others dying off and eating their share I don't see how they can survive investment drying up.
Once the true pricing so they're profitable comes out I suspect there will be a lot less interest in the products on offer barring a breakthrough reducing resources.
Hopefully it becomes a big boon for local models though!
Upvote
13
(14
/
-1)
The Sheep Look Up
Ars Praetorian
Do you believe that document to be in any way legally binding, Miss Whittier?
Upvote
0
(6
/
-6)
I bet they first find a loophole (ads not technically part of the chatbot) and then go back on their word (ads in the chatbot).
But their cash flow and investor sentiment must be such that they can afford to wait for OpenAI to take the blowback for a while.
Remember when Google ran Pixel ads that made fun of the iPhone for ditching the headphone jack. The removed the headphone jack from the next Piexl phone 6 months later.
Upvote
23
(24
/
-1)
J.C. Helios
Ars Scholae Palatinae
Reminds me of that Samsung ad making fun of Apple for removing the headphone jack.
View: https://www.youtube.com/watch?v=wQ4ys-R1B_8
Like, no kidding the dongle life was a downgrade, but Apple did it to make money selling high-margin Bluetooth earbuds... same reason Samsung ate crow and removed the headphone jack shortly thereafter.
Upvote
13
(16
/
-3)
Google search results will never have ads. Amazon searches will never have ads.
They just have "Promoted listings". It's entirely different. Anthropic AI will just have "Promoted mentions".
They just have "Promoted listings". It's entirely different. Anthropic AI will just have "Promoted mentions".
Upvote
8
(10
/
-2)
Meanwhile, my $200 (free and clear, unlocked) Moto G84 has a big battery, plenty of grunt to run everything except the latest 3D games, and still comes with a headphone jack.
Have fun keeping up with the Joneses, guys.
Upvote
-17
(2
/
-19)
I think that Google, with it's years of investment in TPU's will pull ahead in terms of cost per inference. Others are already attempting to do the same and/or licensing Google TPU's. The race will be amongst those capable of achieving cost performant inference more than anything else.
The question isn't can useful work be done, or who has the most useful model, it's who will be able to do so at a sane operating cost in 3-4 years or whenever the circular investment begins to fall apart. That race is already on and showing great progress. Nvidia just aqui-hired a company working on exactly that.
Google has the biggest moat. Meta is quietly building their own. OpenAI is scrambling to do so with Broadcom. Anthropic is licensing TPUs from Google with the largest TPU deal in Google's history.
Upvote
7
(7
/
0)
Resistance
Wise, Aged Ars Veteran
It's amusing how we are using advertisements as evidence in this context.
Upvote
7
(7
/
0)
Oh please. Inb4 they insert ads because where haven't we seen this sentiment before? Especially for an AI company where the seemingly endless ad money might help offset their costs.
Upvote
6
(6
/
0)
fellow human
Ars Praefectus
I dunno if this ad will age well for them but I gotta say they nailed the creepy AI rendered guy, that's pretty good 
Upvote
10
(10
/
0)
I was gonna say "hey, it's this millennium's 'Don't Be Evil'".
Upvote
10
(10
/
0)
YES.
Then maybe people will understand they're not unbiased.No, no, that isn't what it appears. It means famed musician and actress Courtney Love is sharing a password with...her brother. Yes. Probably for an Office360 subscription.
Upvote
-8
(0
/
-8)
Upvote
-1
(2
/
-3)
It is now Netflix's policy to annihilate all love in the world by any means necessary
Upvote
7
(7
/
0)
100% they're just talking their book. They don't have as many free users so they can't monetize meaningfully with ads anyway.
I will say I use Claude Opus 4.5 all day long and it is vastly superior for coding. OpenAI is in serious panic mode for real work and needs to pivot hard into monetizing its userbase and network effect before it loses it to Anthropic or Google.
Upvote
7
(8
/
-1)
If NVIDIA's new Rubin Platform pulls off its 10x reduction in inference costs you will see huge price/cashflow changes in the next year. Google's target for its TPUs is 100-1000x reduction in inference costs.
Upvote
-4
(6
/
-10)
There is only so much that can be done to make dedicated FP16/FP8 math more efficient. Google's TPUs are hilariously efficient in doing exactly that, but it's not like there are orders of magnitude to be had there.
And keep in mind, Google is making TPUs for Google stuff -- they train a huge model once, then infer it a billion times to give you shitty AI "search" results on top of their pages. That's not really where the truly useful stuff, also known as the money, is -- generic models are already hitting walls all over the place. Doing domain-specific things on changeable environments, like code analysis on large codebases, that can be useful, but you can't really do that on a generic model, the context gets out of hand. At some point you're going to have to train, train, train that model on each revision for it to be useful, and Google TPUs can't do that.
And as has been said a million times now, the real fun starts when these companies start charging what their services actually cost. At this point they're just sinking trillions into people becoming dependent on them, but the piper will have to be paid sometime.
To wit, Lee's article about his vibe-coding a syntax coloring Python script should give you a hint: he used day after day after day of "free" Claude credits to do it. I'm willing to bet that the power and amortized hardware cost of those "free" credits might be something like $10/day. Would you still "vibe" as much with doing a random hobby project if it costs you $50 or even $100 for your agent to do it?
Upvote
22
(22
/
0)