[crowcroft]

"Ads that support free access and don’t change ChatGPT answers."

I understand what they're trying to say but this statement is factually incorrect. Answers never used to have ads, and now they do.

In the very first example, if ChatGPT wasn't running ads Heirloom Groceries wouldn't show up, therefore it is a different answer.

OpenAI is splitting hairs and implying that the ad and the 'answer' are two separate components making up a response, but that is not how users will see things, and OpenAI will have ever increasing incentives to blur the two.

[jonas21]

It's correct in the same way as saying ads in the New York Times don't change the articles. Seems fair.

[crowcroft]

I think a better comparison is saying that search ads don't change search results (but it does change the results page).

The point is that the language and nuance ends up being lost on a large portion of the audience.

[thrance]

Funny, because the impact of introducing ads on the editorial line of any publication that does it is very real. I'd expect the same from ChatGPT.

[vdfs]

Just like some youtube content with built-in ads about AI tools while the video is bushing on AI tools

[padjo]

But they do?

[basch]

doesnt it just mean the ad isnt part of the context? that they are isolated from each other and the ad cant steer the conversation?

I get what youre saying, but I do think its important for them to point out the ad is sandboxed.

[crowcroft]

I totally agree, but the framing is critical.

I guess the question is, when I write a prompt into ChatGPT is the answer the entire response I get back, or is the answer just one part of the response I get back.

To date the entire response = the answer and so users likely see them as synonymous. That metaphor is being broken now and we're saying "no actually the response contains multiple things and only one part of it is the 'answer'".

Maybe I'm the one splitting hairs though.