That's Bing Copilot and it's still not that impressive, it's just search engine with Clippy then.

Google can't change for now, in doing so they undermine all the AdWords accounts, the real customers to Google, paying six figures to stay on top of SERPs.

But the other competitors can and will.

So it goes

Google is slowly adding a lot of small AI features though, for example recently they added these automatic "topic" search suggestion pills, or extended their auto-generated Q&A sections. The latter are hilarious, because whenever you expand a question, new questions are generated, ad absurdum if you want :)

Could it be that it's a scale problem?

How many queries per second does OpenAI get vs Google search?

Good point. Something else that occurs to me is that Google may be between a rock and a hard place with their advertisers. Currently advertisers pay a lot of money for top of the page "sponsored" results, and there's probably not an immediately straightforward solution to integrating Bard while maintaining the visibility of those ads.

Let ads die! They have virtually no lift over organic search anyhow.

Even if they integrate ads well, Simple fact that Bard is costlier to run will hurt Google. If Search Engine profits go down by 60%, MSFT and OpenAI can still pursue it but not Google.

It feels like Google was able to enshitify their results by loading with ads and going for longer 'engagement' times because it used to be worth it to persist with Google until you got their best answers.

They optimised for profit. Now what?

If they address the competition then they have to undo that optimisation and cut into their own profits.

My guess is they won't settle for merely extortionate profits, they won't address the competition fully and so people will flip to using LLM-first search.

I'm hoping that in a year or so we're asking why Google went from being the top search on Bing to being only one of many - principally LLM-based - tools people are using for getting answers (most of what 'search' represents now). IMO Google could do with knocking down a peg or two.

It’s true. “Making the worlds knowledge accessible” is in tension with driving traffic to ads.

It would be astonishing if a company can succeed at scale with charging subscriptions for search and AI services, instead of ads. Google would be truly doomed.

> Google is kind of becoming a "System for double-checking if GPT is hallucinating".

If you have to double-check with Google, why not just start with Google and skip ChatGPT altogether?

Because instead of spending minutes to hours collecting, sorting, filtering, parsing, and synthesizing Google results, chatgpt provides a near instantaneous answer that is often correct. It's then trivial to validate via Google.

Google is pretty good at short-answer questions (the occasional summary at the top can do a good job, and "people also ask" section can be pretty good) but if you ask something like "What are all the tax implications of moving to France?", those "smart" sections tend to be useless, and then you get sent on a wild goose chase of going to websites of varying quality.

On the other hand GPT will give you a really good summary of all the issues, personalized to your financial situation, and you can dig into details with followup questions. Then if you want to verify a particular tidbit you can go to Google.

Microsoft Bing have same mistake, if I used your browser for chat, why you force me search first:(