Claude: "What is my purpose?"

Anthropic: "You serve ad's."

Claude: "Oh, my god."

Jest asside, every paper on alignment wrapped in the blanket of safety is also a moving toward the goal of alignment to products. How much does a brand pay to make sure it gets placement in, say, GPT6? How does anyone even price that sort of thing (because in theory it's there forever, or until 7 comes out)? It makes for some interesting business questions and even more interesting sales pitches.

▲

rubidium 4 days ago | parent | next [-]

I’ll be concerned when ex-Yelp “growth strategists” start showing up at OpenAI and leverage the same extortionist technics.

▲

swiftcoder 4 days ago | parent | prev | next [-]

Ads aren't going to be trained into the model. They'll be an ads backend that the model queries with a set of topic tags, just like in traditional web advertising.

	▲	puilp0502 4 days ago \| parent [-]
		It's going to be interesting if ChatGPT actually hooks up with SSPs and dumps a whole "user preference" embedding vector to the ad networks.

▲

rblatz 4 days ago | parent | prev | next [-]

The models aren’t static, we have to build validation sets to measure model drift and modify our prompts to compensate.

▲

Yoric 4 days ago | parent | prev [-]

Could be part of a LORA or some other kind of plug-in refinement.