How ChatGPT serves ads

Less than two years ago, Sam Altman said

> I kind of think of ads as a last resort for us for a business model. I would do it if it meant that was the only way to get everybody in the world access to great services, but if we can find something that doesn't do that, I'd prefer that.

So, is this OpenAI announcing they're strapped for cash?

Abraham Lincoln was the 16th president of the United States of America. He was best known for being “Honest Abe”, writing the Emancipation Proclamation, and playing RAID: Shadow Legends, an immersive online experience with everything you’d expect from a brand new RPG title. It’s got an amazing storyline, awesome 3D graphics, giant boss fights, PVP battles, and hundreds of never before seen champions to collect and customize.

These are the less worrying kind of ads in our future.

Seeing how google has been fighting SEO for ages, what's going to happen when companies figure out how to inject ads into the model?

We haven't yet seen the problem of adversarial content in play, I think.

Imagine people like Sam Altman having access to frontier models without any restrictions that allows them plot strategies to reach their goal in a long term timespan that you don't even realize when it even began.

That's scary. They could fight for censored model for the mass, not for them.

Since they are served as distinct events then I would think they should be easy to block.

Once the ads are injected directly into the main response is when things get interesting.

The ads are in the free tier and the new ad-supported $8/month plan.

Every time this comes up there are comments assuming that ads are being injected into the normal plans, but these are for the free tier and the new Go plan which warns you that it includes ads when you sign up.

"Ads don’t influence responses" - they just arrive in the same payload, measured with four layers of attribution and politely pretend to be coincidences.

Schrodinger’s monetization: completely separate, yet somehow there.

I'd always thought that ChatGPT ads would be indistinguishable from actual content.

It is one of the eternal lessons; All tech business plans eventually lead to serving ads. At least until we ban pixels / 3rd party tracking.

I see OpenAI making a significantly larger amount from defense contracts than from advertisements pumped into chats. So I wonder whose bright idea it was to create a public perception risk.

So news about OpenAI demise is real. They can’t sustain themselves without ads.

The schema is literally named single_advertiser_ad_unit. The single_ prefix is doing all the foreshadowing you need.

Can't wait for "watch this ad for 90s to use xxhigh on your next prompt!"

Can't wait to see how the next election(s) turn out--I'm unsure that a properly well funded campaign would skip the opportunity.

So, we need a lightweight local LLM, that is tuned to remove ads from online LLM results.

Google was built on ads and it wasn't bad for them, its no some tabu forbiden word or business model- as a power users its not for us, but for my mom - it will work

Ads fund the "free" internet. Like it or not, that's the price of the "free" compute. I only hope OpenAI won't enshittify paid offerings just like Anthropic did.

And it begins.

I was looking to see if BZR referred to a 3rd party ad network. I didn't find anything, but apparently someone has replicated OAI's system and you can run insert it into your own LLM.

GH: system32miro/ai-ads-engine

figured this was inevitable once they started the free tier. the attribution loop being a separate event stream is actually kind of clever engineering though -- means they can A/B test ad formats without touching the core model response

Gemini and Copilot are already full of ads, pushing the companies ' own services. I guess the only difference is here that OpenAI has nothing else to push, so they have to use external ads.

If highly targeted/tailored LLM ads on free accounts aren’t good enough for HN, are any ads acceptable?

Let’s be reasonable.

Really well written, technical post. Good read.

I don't like anything about this.

They must be desperate to try to push ads down to people. I am living a mostly ad-free life, e. g. ublock origin and what not, so using something like AdChatGPT would not make any sense. One can sense how the money-flow leads them to try to design a system people depend on - and then they cram down ads into those people. Very unethical.

Here we go again. Imagine if we put as much engineering effort toward actual things that help people, but more ads it is, as always. This is proof AGI doesn’t exist. If it did, it could come up with a better business model than more fucking ads.

Remember that ads are the "last resort" for OpenAI, and they're doing this despite the fact that it's "uniquely unsettling", according to Sam.

Was he lying, or has OpenAI given up hope that this train wreck works economically without enshittification? Neither option is good, but I don't really see a third.

Interesting, no bidding flow entirely first party and contextual.

I've seen chatgpt suggest me more amazon products lately

Not to me they don’t, cause I canceled my account and stopped using their products when they made the announcement.

I don't get what's wrong with charging for your product. Like get rid of the free tier and make a small tier with an easy to serve model for like 5 bucks. Is it still the DAU rage of the 2010ss that's driving burning money?

Let the enshittification commence!

This is gross

It feels like we’ve been in the golden age and the window is coming to a close

Let the enshitification begin, I guess

[flagged]

[flagged]

That's cool, I'll never see them.