[treyhuffine]

The recent success of ChatGPT shows that the current search experience is likely deprecated. With LLMs, you can get information immediately that directly answers your query.

ChatGPT has its limits too though - it's not always right, you don't have options other than to re-ask or regenerate if the one answer doesn't satisfy what you need, we're limited to only text, and the experience may be just a bit too contained/convergent. It feels like it's going to take another product or iteration to really get there.

The question I keep asking - what does the future of problem-solving and finding knowledge on the internet look like in an ideal world given the tools that we have or could now build?

[flir]

> With LLMs, you can get information immediately that directly answers your query.

Not as long as they're relying on humans to write pages with that information in. When nobody visits, what's the point in writing a page?

[twelve40]

I keep looking for a coherent answer to that question, but have never seen one so far, these are unimportant details that are swept under the AI enthusiasm rug.

[flir]

Someone (I wish I could remember who) once argued that if you're trying to create an ecosystem around your product, you can't hoover up all the value, you have to leave some for the little guy. I've felt for a while that Google, with its various attempts to put the answer on the search results page, might be skirting dangerously close to that problem. It feels like the only value they're leaving on the table is the chance to sell stuff.

[twelve40]

so... how much value does ChatGPT as proposed solution leave to the little guy?

[flir]

My expectation is "not much", because if the fact is utterly divorced from the source, there's not even any point writing near-spam pages to attract people to the site where you're selling something. But I think we'd need to see a working implementation to know for sure.