[krat0sprakhar]

> Before, we browsed recipe online or watch them on youtube. Now, we just ask ChatGPT.

> Even small things—if I can still eat those eggs or where to travel next—are now AI-assisted.

Both of these are true for my wife and me as well. It's incredible how valuable these LLMs are for short, quick answers in the day of SEO laden crapware on the internet.

For coding (side-projects), I find CLINE super valuable and use it quite often to get refactoring ideas.

[grok22]

But the answers are really coming from those "SEO laden crapware" sites that ChatGPT was trained on -- in some ways ChatGPT has kinda stolen information and summarized it, but what happens over time? How do arrive at a new recipe for the same old things?

[akomtu]

Google was like that in the early days. In a decade, datasets of authentic texts will be goldmines, carefully guarded and exploited. From those goldmines numerous AI slop-machines will be creating the derivative product: a mixture of 1% gold, 49% filler sugar and 50% ads poison. Obtaining authentic texts will be similar in difficulty to obtaining pure gold bars.