If ads are clearly labeled as "ad" or "sponsored", and they only appear for free users, I think seeing ads is a pretty reasonable price to pay for those who want to use the service for free.
If they're not labeled, or are shown even to paying users, I think that's a problem.
All ads start as clearly labeled and distinctive. Then via the magic of iteration and A/B testing they magically evolve to become visually indistinguishable from the rest of the content except for what’s required by law.
They'll eventually want to set it up so you read the sponsored content first, before seeing the tag saying it's an ad. You're more likely to absorb it then.
Especially if it's LLM-generated to fit with the context, the message will slip right into the mind. Then a little "(Sponsored)" at the bottom after you've already consumed the ad.
This is a bit like how ads are presented on X, they look like regular posts or replies but they usually feel off topic and you're thinking "huh, this doesn't fit the discussion". But LLMs will allow much more seamless and sneaky ads.
The iron law of encrapification: if a company can make more money by downgrading the user experience, it will. I imagine within Apple there were still people who advocated for a better, more transparent user experience, but ultimately they seem to have lost out to services people who just want to grab more money.
It's unfortunate because user experience was a core differentiating advantage for Apple that got them to where they are now.
People who do not pay for ChatGPT often have money and prefer not to pay for for a subscription for several reasons including, but not exclusively:
1) They don't use ChatGPT often enough to justify it
2) They use alternatives primarily (a subset of #1)
3) They choose to spend their money on other things
How can an advertiser tell the difference? Which is a stronger signal of having money: paying for something, or not paying for something? Furthermore, with all those reasons, why would advertisers prefer those people in ChatGPT? Advertisers are trying to change your behavior, usually to spend money the way they want you to. If you’re rarely using the service and don’t easily part with money, you’re probably less worth persuing than… well the person who is the opposite of those things.
Ads will lower the quality of the training data, an RAG is more likely. Pay to get your product's INSTALLME.md ranked under some specific semantic vectors.
Every other iteration of a service that introduces a free ad-paid tier then ratchets it to bifurcation of premium in to 'premium' and 'premium with no ads' and then on and on.
My bet is that there will be ads for both groups. The paid group is arguably more valuable from an advertiser’s standpoint, and you can target heavy users with more granularity.
I’d rather they served ads. The economy is somewhat broken right now, with the way these things are bypassing all regular information channels. This will hopefully create lots and lots of new business for 3rd parties again.
Ideally, they’ll introduce a whole new level of targeting relevance, which will be good for both advertisers and prospects.
we’re talking about different things. there’s meritocratic fairness where producers are paid fairly for their work, and there’s a functioning economy, where there are simply enough economic opportunities to sustain established norms of commercial participation by the broad population.
Thinking about the power and reach of political ads served by social media companies over the past 10 years, this is gonna be a whole nother bucket of worms.
I just canceled my $200/month GPT-Pro subscription. 5.2-Pro is in decline -- it has been getting noticeably worse at a steady rate since introduction. At this point, it's not appreciably better for most queries than Claude 4.5 Opus, and Opus is roughly 10x faster.
Noticed the same. Also, I noticed the same with the prev version, immediately improved when switching to the new 5.2.
The smoking gun is the time. If I ask it a question that's subtle in "thinking" mode and it starts replying in a few seconds, the answer will probably be trash. I'm almost sure they degrade the models over time.
An interesting approach - start strong when everyone is running benchmarks, taper off through the life of the release, introduce new version that is no better than the last version but it magically seems much better by comparison to the degraded previous version. Rinse, repeat, grift.
My impression is that recently ChatGPT tries to avoid
going out to research on the Internet as long as it can. I have to tell it to pull info from the web or verify its answers on the web explicitly.
My non-existent marketing instinct would tell me that they are trying to keep you inside the app to convince you that ChatGPT is the internet, the same way some people wouldn't know there's life outside Facebook.
My grumpy instinct tells me they know that they're poisoning the internet and they have given out on trying to weed out the fake websites from the real ones.
I’d be curious what proportion of their usage is on the mobile app and doesn’t need to worry about a significant number of users having adblockers. My instinct would be probably a decent majority is mobile, but not as high as something like Facebook, but that’s just a guess.
How will you know your response doesn't contain an ad?
I see it as the responses eventually mimicking all of the marketing spam posts, where company Y compares it's competitors poorly or does a thought leadership piece on how you can do X by hand or have their product do X for you.
> In a science fiction story, if you invented a superintelligent robot and asked it how to make money, it might come up with cool never-before-seen ideas, or at least massive fun market manipulation. But in real life, if you train a large language model on the internet and ask it how to make money, it will say “advertising, affiliate shopping links and porn.” That’s the lesson the internet teaches!
But I think it makes a lot of sense for very popular consumer products. In my honest opinion, I much prefer having services like Google, Youtube, Gmail, Maps, ChatGPT etc exist for free, but with ads, rather than not exist at all. Preferably with an option to pay and remove ads
Nowadays I'm happy to pay for Youtube premium or LLM, but back during my student days I could not really afford it - and I'm glad there was a free tier (with ads)
>In my honest opinion, I much prefer having services like Google, Youtube, Gmail, Maps, ChatGPT
I don't use any of these except YouTube (if only I could find the content elsewhere…) and I still pay for them when I purchase anything advertised on these properties because, of course, the companies advertising on Google makes all their customers pay for the free (lol) services. All advertising expenses are included in the price of the products, even if you never saw any ads.
We could easily charge for each of these services and still have them. Advertising is not necessary at all. It's just a way to make others pay for your services. It's a free riding problem to externalize costs on those who don't partake in the scheme.
Pay your share and don't call free what others will subsidize. Unless if a public service and we collectively agree on the split (vote and taxes, which we can debate publicly)
Right. But a good portion of the world can't afford the premium and having access to these services is still valuable. For every broke student or someone from a poor background, who probably don't make any money for the company (due to not buying advertised stuff), there's someone from a well off background, who will more than subsidize it by virtue of clicking on a lawyer ad (or whatever)
Nowadays I'm happy to pay, but that wasn't always the case. And I personally think that having an ad tier and fee tier is fine. Serves everyone
As a general rule of thumb in sites like Reddit and HN - the quality of votes is significantly lower than the quality of comments. This is because it takes much more effort to comment, so there is a selection bias.
I'm not convinced that downvotes add much value. They should be a "this is irrelevant/spam" button but in practice they seem to be used as a "dislike" button to enforce groupthink.
Yeah… I don’t understand how anyone could look at the prevalence of advertising and affiliate links on the internet and believe they would for some reason stay away from the LLM products.
Sure Sam Altman and his $200/mo subscribers won’t see them, but it was clear they were coming for free users.
Voting can be for all sorts of reasons. Sometimes I downvote things like that because I want to bury bait that would send an argument into a well-trodden and boring direction.
Good thing I didn't develop an unnecessary dependency on this product, so now I don't have to suffer through its enshittification. It's almost like it was obvious this was going to happen years ago.
By the way don't sleep on this detail:
> The banner ads will appear in the coming weeks for logged-in users of the free version of ChatGPT *as well as the new $8 per month ChatGPT Go plan*
Even if you pay for the product, you're still the product. If we don't own our software, our software will own us.
It's not about how it starts. It always starts small and measured, but once you open up to ads, you open up the pandora's box of enshittification paths.
I have made some purchasing decisions on expensive products based on analysis ChatGPT has done for me, if they can capitalize on that, it could be a decent way to make some money, as long as they remain unbiased and basically just function as an affiliate marketer. Sometimes I do want to be sold on something.
What is the point for a company to pay OpenAI for products that it would recommend anyway? Companies are going to pay only if they can add bias to the results otherwise there is no point.
Many of the advertising targets (you) will have confessed or indirectly revealed many of their aspirations, interests, hobbies, health, life and relationship problems and preferences in “chat”
In a way they can’t get due to increased use of ad blockers or tightening restrictions on data brokers (California and EU GPDR) etc.
So it’ll be very competitive for an advertiser to go with your ai “friend” who knows all about your hemmorhoids, booze and sex problems. All of which Google and Meta can infer or at least pin on you via guilt by association.
Meta screwed that one by breaking up known connections and communities and putting AI slop and promoted content front and center . They can infer less from who you know or interact with because they stopped caring about connecting you to real humans you actually know, years ago.
Banning all your friends and breaking up all those core groups for voting wrong or thinking wrong or whatever more closely suited their interests and agendas at the time.
It might know what I do for work or living based on what I ask for help with etc.
Our approach to advertising - https://news.ycombinator.com/item?id=46649577 - Jan 2026 (227 comments)
If they're not labeled, or are shown even to paying users, I think that's a problem.
Especially if it's LLM-generated to fit with the context, the message will slip right into the mind. Then a little "(Sponsored)" at the bottom after you've already consumed the ad.
This is a bit like how ads are presented on X, they look like regular posts or replies but they usually feel off topic and you're thinking "huh, this doesn't fit the discussion". But LLMs will allow much more seamless and sneaky ads.
https://9to5mac.com/2026/01/16/iphone-apple-app-store-search...
It's unfortunate because user experience was a core differentiating advantage for Apple that got them to where they are now.
OK, maybe not all shareholders are playing the short game, but I feel like a lot of them are.
Why would advertisers prefer people without money to people with money?
People who do not pay for ChatGPT often have money and prefer not to pay for for a subscription for several reasons including, but not exclusively: 1) They don't use ChatGPT often enough to justify it 2) They use alternatives primarily (a subset of #1) 3) They choose to spend their money on other things
Btw, the end game is probably having ads in the llm context .... or directly in the llm training set.
Also, everyone gets a free pony.
Ideally, they’ll introduce a whole new level of targeting relevance, which will be good for both advertisers and prospects.
I just canceled my $200/month GPT-Pro subscription. 5.2-Pro is in decline -- it has been getting noticeably worse at a steady rate since introduction. At this point, it's not appreciably better for most queries than Claude 4.5 Opus, and Opus is roughly 10x faster.
The smoking gun is the time. If I ask it a question that's subtle in "thinking" mode and it starts replying in a few seconds, the answer will probably be trash. I'm almost sure they degrade the models over time.
Could it be that they are trying to save traffic?
My grumpy instinct tells me they know that they're poisoning the internet and they have given out on trying to weed out the fake websites from the real ones.
Or to put it another way, I'll be interested to see how long before the ads become inseparable from the actual content of the response.
I see it as the responses eventually mimicking all of the marketing spam posts, where company Y compares it's competitors poorly or does a thought leadership piece on how you can do X by hand or have their product do X for you.
> In a science fiction story, if you invented a superintelligent robot and asked it how to make money, it might come up with cool never-before-seen ideas, or at least massive fun market manipulation. But in real life, if you train a large language model on the internet and ask it how to make money, it will say “advertising, affiliate shopping links and porn.” That’s the lesson the internet teaches!
But I think it makes a lot of sense for very popular consumer products. In my honest opinion, I much prefer having services like Google, Youtube, Gmail, Maps, ChatGPT etc exist for free, but with ads, rather than not exist at all. Preferably with an option to pay and remove ads
Nowadays I'm happy to pay for Youtube premium or LLM, but back during my student days I could not really afford it - and I'm glad there was a free tier (with ads)
I don't use any of these except YouTube (if only I could find the content elsewhere…) and I still pay for them when I purchase anything advertised on these properties because, of course, the companies advertising on Google makes all their customers pay for the free (lol) services. All advertising expenses are included in the price of the products, even if you never saw any ads.
We could easily charge for each of these services and still have them. Advertising is not necessary at all. It's just a way to make others pay for your services. It's a free riding problem to externalize costs on those who don't partake in the scheme.
Pay your share and don't call free what others will subsidize. Unless if a public service and we collectively agree on the split (vote and taxes, which we can debate publicly)
Nowadays I'm happy to pay, but that wasn't always the case. And I personally think that having an ad tier and fee tier is fine. Serves everyone
I struggle to understand people getting butt hurt about a free service showing its users adverts, that will keep the service free.
They should have done this earlier, so their adds would be better by now, and they have a better chance against Google.
Sure Sam Altman and his $200/mo subscribers won’t see them, but it was clear they were coming for free users.
Yet. Amazon Prime has ads despite it being a paid service.
By the way don't sleep on this detail:
> The banner ads will appear in the coming weeks for logged-in users of the free version of ChatGPT *as well as the new $8 per month ChatGPT Go plan*
Even if you pay for the product, you're still the product. If we don't own our software, our software will own us.
In a way they can’t get due to increased use of ad blockers or tightening restrictions on data brokers (California and EU GPDR) etc.
So it’ll be very competitive for an advertiser to go with your ai “friend” who knows all about your hemmorhoids, booze and sex problems. All of which Google and Meta can infer or at least pin on you via guilt by association.
Meta screwed that one by breaking up known connections and communities and putting AI slop and promoted content front and center . They can infer less from who you know or interact with because they stopped caring about connecting you to real humans you actually know, years ago.
Banning all your friends and breaking up all those core groups for voting wrong or thinking wrong or whatever more closely suited their interests and agendas at the time.
It might know what I do for work or living based on what I ask for help with etc.