Adapting to our new AI overlords

Following my “SEO is dead? Again?” post, I’ve stimulated my appetite to explore further. And disagree with myself.

With hindsight, I may have been too eager to paint a binary picture, where either everything stays as it is today or no one uses Google any more.

One of the reasons for this error is that I believe we’ve become conditioned to accept what we have as the norm.

Google doesn’t define the norms forever.

Before Google, there were many search engines to choose from. AltaVista, Lycos, AskJeeves, Yahoo and more. But none seemed to dominate, as some were better at certain things, but none consistently excelled.

I remember a tool called WebFerret that combined the top results from all the main engines into a diverse and refined list. That’s how poor the results were at the time.

Then Google came along and gave people what they were looking for. Without the animated and annoying ads that some of their competition were using.

There’s irony there.

Over time, Google grew into its ecosystem and became synonymous with the web. Not just searching the web, but the web itself.

To this day, a significant number of people use Google without even realising it.

An intervention is needed, to see the reality.

It’s the online equivalent of Stockholm Syndrome.

We don’t like what Google does with our data, nor how they decide what to show us.

Yet we’ve come to accept and even like how they do it.

We’ve become used to their norms:

Search and click your way through the results until you find what you’re looking for.

Or search, refine your search, and then click your way through the results.

For all the efforts that Google have made over the years, the only real impact has been to sometimes reduce the number of clicks to find the answers.

The methodology and experience remained the same.

Search, click, sift, refine and identify.

Repeat.

Then repeat again.

This has been the foundation of web navigation for decades.

One of Google’s greatest achievements was making us think this was good enough.

There may be a better way.

In hindsight, it’s obvious that when people search on Google, they’re not looking for links, they’re looking for answers.

And it’s fascinating that for me, once I’d experienced a new way of finding answers, I didn’t want to go back .

But there’s a tradeoff here.

Even though the experience is quicker, it can be harder to gauge accuracy.

On Google, when you see a long list of links suggesting the same answer, there’s a level of reassurance in the wisdom of crowds.

But when you get one single answer, you either have to assume it’s correct or go to verify it for yourself.

This is even more complex when questions are asked with no definitive answer, such as personal preference or opinion.

For example, I searched for “Who is thought to be the greatest singer of all time, by whom, and why?” in ChatGPT, Perplexity, Google and Bing.

I’ll scroll to where the actual answers begin to avoid interface and presentation issues, and the screenshot will include whatever fits in the browser without scrolling further. I’ve also marked the number of answers for each.

ChatGPT:

Perplexity:

Google:

Bing:

I’m going to overlook the sheer madness of Freddie Mercury not being the first result for every search.

Aside from that, what surprised me was that they all had the same number of answers within the browser window.

But with interesting differences.

ChatGPT had the clearest layout and most comprehensive answers.

Perplexity had the best summarised and easiest to absorb information.

Google had an astonishing amount of information, and an almost overwhelming number of links to click on.

And Bing had a surprisingly decent presentation of information, with far fewer links than Google.

This seems like a throwback to the old days — of different engines being better at different things.

Which provides the better results?

It depends on what the searcher is hoping to find.

ChatGPT reminds me of the British phenomena of the Pub Expert — the man who is an expert at everything, but never has to support his opinions with facts.

Perplexity, on the other hand, are taking the approach of sticking to facts, but with links for you to explore the sources for yourself.

Bing did a surprisingly decent job of presenting not only the facts, but also some good snippets and links.

Google surprised me the most. When compared to the others, it simply looked like a mess.

What about factual information?

Using the same approach as above, I searched for “what was the hottest recorded temperature on earth?”

ChatGPT:

Perplexity:

Google:

Bing:

Again, while there were differences, the similarities were striking.

All gave me the same answer. It’s reassuring to see that there are still some generally accepted facts.

ChatGPT was succinct, while giving all the information you might want to see.

Perplexity gave more details and links to their sources.

Google gave all the information we might want to see, followed by plenty of links, of course.

Bing gave the least detailed summary and many links again.

So what to make of this?

Searching for facts highlights the strengths and weaknesses of each platform.

But I am most struck by something I overlooked in my previous post.

Even today, there is no single best platform. It depends on the needs of the searcher, their search at that moment, and their personal preferences.

So I’m going to refine my original prediction.

I don’t think that Google should be preparing for their demise. Far from it.

That said, if they continue their strategy of serving up answers in the same way that they have for 20 years, their days are numbered.

Eventually.

Why? Because better ways of presenting information are now being shown to people.

And we might find ourselves unwilling to tolerate the old ways of doing things.

However slowly they may appear to start, such is the nature of most revolutions.

Who’s going to be first to dominate?

I don’t know.

But imagine this.

Imagine a version of Perplexity with a greatly-improved design.

Instead of offering you a basic “Ask anything” prompt, it would ask you what you were looking for and how you wanted to see it.

If the design allowed me to not only ask my question, but also indicate how I want the answers presented, that could be a significant advantage.

Whoever resolves this issue will have a very real chance to dominate the landscape, and possibly the world.

What should you be doing?

The previous advice of extreme proactivity still applies. But I would also recommend a little experiment.

Type the following into your AI engine of choice:

[your company] – who are they and are they good and trusted

This is what we found when we tried this on ourselves:

This amazed me a little, but I’ll allow you to draw your own conclusions.

I’d also recommend not burning any bridges just yet.

We don’t know if or when Google will become obsolete. More importantly, we don’t know how long this will take.