A 'Litmus paper' hypothesis

I have a hypothesis about Google’s push to insert AI into Search, effectively killing the thing as we know it (of course, they cannot say it’s a Large Language Model; no, they will call it AI with a stubbornness befitting of a greater cause). And because I have some time to spare, I will share my hypothesis here so that later on I can look back on this moment and pat myself on the back (either as a sign of approval or to help myself come to terms with a mistake I made).

The KPI of Progress As you might have seen on the Internet, people didn’t take kindly to the unveiling of the newest plan for the Google browser, probably because what is viewed by the investors as a measure of success doesn’t translate at all to the language of normal human beings.

In the keynote video from the Google I/O 2026(https://io.google/2026/explore/google-keynote-1) conference, Sundar Pichai (CEO of Alphabet) shows a graph of how ‘the monthly usage of tokens across our surfaces’ grew from May 2024 to May 2026, with a big bold ‘7x of Y/Y growth’ and a blue line climbing straight into the sky of limitless possibilities (I presume). This is the measure of success. The KPI of progress. This number that is shown to investors tells a story of success, of the advent of the new era, the era of AI. [I must say that I don’t remember the last time when the advent of anything in the tech industry took so much time… We have been in this new world for a couple of years now, and we have yet to see anything useful and groundbreaking. I mean, even the ground under the data centres that are supposedly on their way and under construction, in many cases, hasn’t been broken yet.]

‘Over the past 12 months, over 375 customers have each processed 1T+ of tokens, representing incredible demand for AI across [the tech] industry.’ - according to Sandar.

Over 375 customers - that’s a specific number. Why is there a ‘+’ sign on the slide? Is it 376 customers? Then why not just say the number? Is the ‘+’ there to mean that the number keeps growing, that the progress is happening right at this moment? You should quickly join the crew before it’s too late and the raft sails without you on board, is that it? He also said that 5 of their products have more than 3 billion users each, though he didn’t specify how they arrived at this number. If I have a Gmail account and a YouTube channel, does it count as one entry? Is it only me who is stuck on the numbers, trying to make sense of them? Not sure, but everyone in the audience clapped and cheered, so it probably isn’t that important. The numbers are huge, and that’s enough.

‘It all starts with Search […]. AI Overviews in Search now has 2.5 billion monthly active users. And AI mode has been a revelation, our biggest upgrade in the Search ever. People love it. In just a year, it already surpassed 1 billion monthly active users. […] Today, more than 50 billion images have been generated with our Nano Banana models.’

Don’t get me wrong. I am sure that what he describes is correct, though to me it sounds like hell on Earth. Still, is 2.5 billion monthly active users for a platform like Google really that big? It's not in trillions like anything else these days. It no longer seems that impressive to me. You see, that's the trap of forever chasing the numbers. They lose meaning along the way. And the number clearly doesn’t take into account the generation of my parents. Do they even know they are using this AI Overview? Is it a conscious decision that they are making because they see an added value in it, or is it just because this thing sits at the top of the Search with no option to turn it off? If it’s the latter, then what exactly is represented by these numbers he is quoting? Is it the number of people who are held hostage by the fact that they are not tech-savvy enough to save themselves from AI Overview? Is that a measure of success?

Then he makes a joke by reading aloud what he describes as a real question asked by a person while using Ask Maps (I might be more similar to the older generations than I would like to admit, since I have never noticed that this option was there). Again, everyone laughs, though I must say I find it creepy to think that my plight might have been read aloud in front of a huge audience. And the fact that we don’t get to know this person’s name doesn’t make me feel any more comfortable about the way privacy is treated by the greatest harvester of human-related data that Alphabet surely is. But enough about the conference, the picture is clear now - LLMs will be the Search.

Here we arrive at my hypothesis. It’s simple and elegant if I do say so myself. I propose that we use this ‘Search to LLM’ conundrum as a litmus paper.

Historically speaking, Google’s modus operandi used to be like this:

• If there’s a backlash, they will move away from the idea, put it on hold, and try again in some time, meanwhile making sure that there is enough talk about it that people will slowly get used to it being a thing. That’s what they did with Google Glass, which was once very controversial, but is now returning on a shiny plate of LLMs.

So what I propose is to wait and see: if in the coming weeks Google does not crumble under the backlash, it doesn’t necessarily mean that they don’t care about their customers (although it should be obvious by now that they don’t), it means that they are not able to back out, because they invested so much in this endeavour that they have no choice but to die on this hill. Which, if you think about it, might be the best outcome of the whole LLM debacle.