> slowly they'll become more part of the text

Wouldn't that be quite challenging in terms of engineering? Given these people have been chasing AGI it would be a considerable distraction to pivot into hacking into the guts of the output to dynamically push particular product. Furthermore it would degrade their product. Furthermore you could likely keep prodding the LLM to then diss the product being advertised, especially given many products advertised are not necessarily the best on the market (which is why the money is spent on marketing instead of R&D or process).

Even if you manage to successfully bodge the output, it creates the desire for traffic to migrate to less corrupted LLMs.

▲

the_pwner224 41 minutes ago | parent | next [-]

You could run a second lightweight model to inject ads (as minor tweaks) into the output of the primary powerful model.

▲

barbazoo 4 hours ago | parent | prev | next [-]

I’m assuming they have much more control during training and at runtime than us with our prompts. They’ll bake in whatever the person with the checkbook says to.

	▲	Quarrelsome 4 hours ago \| parent [-]
		if they want dynamic pricing like adwords then its going to be a little challenging. While I appreciate its probably viable and they employ very clever people there's nothing like doing two things that are basically diametrically opposed at the same time. The LLM wants to give you what _should_ be the answer, but the winner of the ad word wants something else. There's a conflict there that I'd imagine might be quite challenging to debug.

▲

likium 3 hours ago | parent | prev [-]

Supposedly Google made their own results worse to improve ad revenue. And I don't see mass migration over to Kagi or Bing.

	▲	ares623 3 hours ago \| parent [-]
		It was still better than those two. Until chatgpt came out and then it was crisis mode.