AI Chat vs Search: The Future Of The Internet

The fight to help you search the web was one of the first high stakes internet battlegrounds. It’s a simple premise: you type in what you want to see, and the search engine finds it for you. Household names like Lycos, Alta Vista, Ask Jeeves made headway before Google eventually became the starting point for most internet visits. 

Search is such valuable ground (both strategically and profitably) that Google was able to leverage dominance there to also own chunks of email, mapping, and video. They even moved further and further upstream into whole operating systems, smartphones, and now dominating the browser market. 

In short: whichever company owns search can shape the whole technology industry. And until now that’s been one company. Any business nerd can reel off many reasons for Google being untouchable. They’ve got the power of brand and top of mind awareness. They have economies of scale for compute power. They’ve got data network effects (the more you search, the better their search can get). They’ve got the power of being the default, both through paid partnerships (like with Apple), and user habits. People would just “Google it”. 

Their lead is so powerful and defensible, that a US court last year ruled that they had a monopolistic position and many expected Google to have to give up Chrome and Android. 

But over the last few years, the unthinkable has started to happen. 

Google’s search is gradually being replaced. Not by a better search engine, but by a completely different paradigm. Chat interfaces, popularised by ChatGPT, let you search for things that you simply can’t on Google. 

We’re no longer limited by “type in what you want to see and it will find it”. You can give extra context, nuance, and images. You can back-and-forth to find the right thing. It can go deeper. Simply put: chat search is smarter. 

A few weeks ago I could hear a lot of noise coming from the city I live near. It was the middle of the week so I didn’t know what it was. I Googled it, but couldn’t find anything. I asked ChatGPT, it thought for a bit, and told me there was a big university derby rugby match on. Bingo. 

A whole lot of internet use-cases that used to be no-brainer Google searches are suddenly better served by ChatGPT. And slowly but surely users are moving over. 

Credit: FirstPageSage 

It’s fair to say that no one knows where this will end up. Can ChatGPT sustain its free model? Will Google putting chat in their main search have an impact? Will people realise hallucinations are annoying? Will Google’s default behaviour win out? There’s no way to know. 

But there is one battle in the future that feels more predictable: the one between chat interfaces, generic search (Google), and specialist search (e.g. Zillow, Pinterest, Amazon). 

To predict the future, we need to look at where we are today. Generalised search has some surprising limitations. There are some specialist search websites that are popular because they do a better job than Google. Zillow lets you see homes and all their information on a map. Pinterest lets you find images, and other images like them. Amazon lets you find the cheapest items for sale right now. 

Why do these still exist in the world of Google? Why didn’t Google Property Search, Google Shopping, or Google Images kill them? The following is what all of those search experiences lack: 

1. A specialised interface Google search is simple, which was the reason for its success. But if you need to show extra information (like on homes), or have unique interaction patterns (like on images) then Google can’t do that. Obviously they could, but it’s hard for them, and users wouldn’t like the inconsistency from the Google they’re used to.
2. Non-linear search. You research a thing, then find the cheapest. You research an area, then find the homes.
3. Extra context for the search. Whether initially, or back and forth. Basically search is one-and-done, and that one is short. 
4. Unique resource. Open Table knows which tables are free. Amazon can change it’s prices algorithmically, on the fly. 
5. Discovery. Sometimes you don’t know what you want. Maybe you want to browse for different toys, or just scroll through homes. 
6. High risk. High risk things are better served by someone you can trust. Amazon has a track record of having things arrive. Deferring to a specialist feels more sensible.
7. Optimisation function. Search is designed to get you the best answer. When buying a home, you want to make the best decision. You want to have seen or considered every home. It’s high risk and extremely non-fungible. “Good enough” vs “the best”. I will look for a long time to get the best house. 

Chat will take over generalised search in areas where getting the right answer takes more than an obvious set of keywords to get a result. Anything where the information is slightly hidden, like finding out the reason for traffic on a given day, or the real story behind a statistic. Both these are possible to find in Google, but you need to leave Google. 

Chat will take over specialised search in areas where there is a a lot of useful extra information. Maybe you want to search for a guitar with exact specifications and has celebrity endorsements. Or you want to buy a chocolate bar that doesn’t have any palm oil. These are hard to search for, but chat can do extra research for you. 

The specialised search websites can add chat functionality, if they also have all this extra useful information. But where they can’t, or can’t adapt fast enough, chat sites like ChatGPT may win. 

And what about basic searches? Will people still go to Google? That all depends on if Google can react and adapt quickly enough. ChatGPT of the future may not only be able to find what you’re looking for, but also research it and buy it for you. The extra functionality that is constantly being added there may see it usurp Google not just as the default place to search, but the default way to use the internet.