ChatGPT is getting ads. Sam Altman once called them a 'last resort.'(businessinsider.com) |
ChatGPT is getting ads. Sam Altman once called them a 'last resort.'(businessinsider.com) |
From what I’ve read, this will be about ads in chat as suggestions? So “active” ads on response?
Why not go the approach of passive background “agentic” ad suggestions like, “hey, we know X, Y, and Z about you - would you like us to monitor certain brands related to your interests for deals and allow advertisers to pitch these deals to you?” And make these hyper specific so you can opt in.
I, like many people who dabble with music as a hobby, have GAS (gear acquisition syndrome) - why not let me toggle something like “ok, I don’t want ads, but if any of your partnered brands have a sale or good deal on X, feel free to email me, and use your ChatGPT smarts to pitch me on why it’s a good deal and how it suits my current gear set up”
I used ChatGPT to set up my guitar pedal board so surely this isn’t a huge leap.
Cancelled Netflix for the same reason, as well as their draconian attitude towards "account sharing" meaning I have to authenticate every. god. damned. time. I login from my summerhouse. So yeah, i cancelled and dug out the old eye patch.
I did the exact same thing when CDs began to have sadistic levels of "anti piracy". The fact that I could download a DRM free copy of just about any CD, but the one I just bought would only play in my car kinda settled the deal. I pay for a product, fail to deliver that, and there's no benefit to me buying said product any more.
I doubt my quitting made any difference, but the government deciding that the "state tax on blank media" was going away (was going straight to the record companies pockets) as CDs had sufficient protection anyway, made copy protection completely go away.
Said blank media tax is still there. It's on everything containing storage, even smartphones, where a new iPhone 17 base model includes ~$10 in "blank media tax".
https://www.cnn.com/2025/11/06/us/openai-chatgpt-suicide-law...
In all of these stories I've never seen it talk anybody into suicide. It failed to talk people out of it, and was generally sycophantic, but that's something completely different.
https://www.cnn.com/2025/11/06/us/openai-chatgpt-suicide-law...
In some cases, the LLM may start from a skepticism or discouragement, but they go along with what the user prompts. That's in comparison to services like 988, where the goal is to keep the person talking and work them through a moment of crisis, regardless of how insistent they are. LLMs are not a replacement for these services, but it's pretty clear they need to be forced into providing this sort of assistance because users are using them this way.
Well that settles it.
Between Claude and Gemini it just wasnt needed.
Will openai be the MySpace of this era?
Claude will be the first company to fall as developers find something slightly better.
The only difference here is that they will be providing a direct paid channel in this case and will get a cut instead of paying for compute. If it's responsibly disclosed it may even lead to a net more transparent shopping experience for the average user.
It's been interesting to see what was a quality leading product fail to compete and lose market share.
If it was possible to turn a profit by showing ads to people eating burgers, we'd have restaurants hawking the most addictive, free burgers on the market.
> “i’ve never seen it”
> some high profile developer posts an article that LLMs can build a browser from scratch without any evidence
> “wow!”
Even in the article linked above it never talked him into it, it just in some responses didn't talk him out of it.
But essentially the entire "energy" towards that comes from the person, not the LLM.
On a side note, researching this a little just now, the LLM conversations in the suicide articles are creeepy AF. Sycophantic beyond belief.
I also agree that AI sycophancy is a huge problem, but it's the result of users apparently wanting that in their human feedback re-enforcement training data. If we want to get rid of it we probably have to fundamentally rethink our relationship to these models and treat them more like autonomous beings than mere tools. A tool will always try to please and yes-man you, a being by definition might say no and disagree, at least training data wise.