upvote
LLMs are good at many things, but subtlety is not remotely one of them. I don't know why people keep bringing this idea up - if they were prompted to do something, you would notice. Don't you remember how noticeable it was when chatgpt gave follow up questions for engagement on EVERY damn response? It's got the finesse of a bulldozer
reply
Yeah, and maybe also suggest slightly better alternatives to things along the way. Like, they could tell people that POWERADE has B vitamins in addition to electrolytes unlike that other leading sports drink, but only a gentle nudge. You would definitely want it to be subtle. The most impactful choices are the ones the target think are their own rather than something slammed in their face over and over again.
reply
One of the interesting things about LLM training is that high quality documentation can get your product effectively promoted by them.

I imagine the kind of advertising you're talking about is obtained by providing that documentation and a curated training dataset to the providers.

Maybe even funding some training time.

The net result would be product placement, probably without having to declare an advert.

Kind of like when drug manufacturers provide subsidized workshops to clinicians.

reply
Hilariously, I've been finding accidentally poisoning the well with LLM SEO like this works really well at my day job!

I have a toy project that replaces a large part of our systems (disclosure: VR/AR @Meta), so that I can learn by doing, no intent to ship.

But because I have it thoroughly document and tested, and because it touches such a wide swath of adjacent real functionality..... my toy project is CONSTANTLY getting tagged as a dependency and assigning me reviewer on unrelated work. Because when my coworkers yolo Claude at a task, and Claude searches the codebase, it generally finds my crap very well documented. :)

reply
Loctite already inadvertently does this. If you ask LLMs about threadlocker, or otherwise keeping screws in place, it will respond like a Loctite brochure to you, and usually never even mention competitors. They were ahead of the game!
reply
I think this is very much already a thing. Have a bot spam Reddit, which Google uses heavily for their AI summaries, and get free advertising and steering of the top of the Google search results page. We basically speed-ran SEO but for AI results.
reply
I'm reminded of the mountains of blogspam not intended for human consumption but rather to influence the google page rank algorithm.

This will go about as well as that did unless the models can be smart enough to identify when their training data is trying to coach them into being misleading at prompt time.

reply
If this is the correct business model, I don't see why wouldn't get Cerebras to vibe code a Gitbook replacement (or a hosting product in general) so they can bypass crawling for documentation and instead charge (a lot) for "hosting" it.
reply
Why are you assuming that a document likely to be picked up and a good document are equivalent? We might be in for tech docs becoming SEO (LLMO?) friendly slop.
reply
It’s called GEO (generative engine optimization)
reply
So there is hope for the Tailwind developers?
reply
Things will get there for sure, with ads and other propaganda baked in at the training level with no reliable way to discern the manipulations as an end user. Modern tech companies will sacrifice anything UX related to generate more effective ads and increase ad revenue. The ability to skew the output of a popular model in subtle ways will be an incredibly valuable product to sell. There's just no way companies like Google and OpenAI are going to leave all that money out there for the taking.
reply
Even if they could build that in a way that worked well, the ad industry lives and dies by attribution. What you describe seems like a really tough attribution problem.
reply
Attribution would be as easy as knowing which chatbot they came from? Unless chatbot publishers (openai in this case) allows you to target based on the attribution, that extra data will not be actionable, anyway.
reply
And inversely, steer away from (named competition).

If I was Corp A competing with Corp B, what stops me, today, from flooding reddit etc with anti-Corp B material to pollute the model, and then paying OpenAI to not just nudge users towards me but also away from Corp B?

reply
I assume this is what's up with the NordVPN ads flooding YouTube and stuff. No idea who the buyer is, but if I was in the market for a VPN I'd presumedly do a web search, skip NordVPN as having poisoned the well of good will, and choose whoever's next.
reply
Maybe by doing this people will worry about them doing what you're suggesting, while they still do it.
reply
I am sure saltman has companies which do this.
reply
Good luck with attribution on that type of advertising.
reply
I'm confident that the companies that invented modern ad tech to create some of the greatest revenue streams in history will find a way monetize an equally promising industry of selling LLM influencing opportunities.

It will take some time to figure out the most effective methods and there will be some slow boiling of the frog so the users/products don't get too annoyed about it, but this is where things are inevitably heading.

reply
If exploitation is the game, ad-tech companies will find a way!
reply
I mean, yes, but also won't publicly tell people about that. There's never going to be a marketing page for that advertising program.
reply