Billions of AI users…?
Between Meta announcing that its AI, Meta AI, reached 1 billion users[1] and Google saying that AI Overviews are used by 1.5 billion[2], I’m curious to know how many of these people intentionally use the feature, or prefer it to what the AI replaces.
AI Overviews appear at the top of searches, with no option to turn them off. Meta AI, I suspect many people trigger accidentally by tapping that horrible button in WhatsApp, in search results across its three core apps, or when trying to tag someone in a group by typing an @ symbol.
It’s very easy to reach enormous numbers when you already have a giant platform. I don’t think that’s even part of the discussion. The issue is trumpeting these numbers as if they were earned, rather than imposed.
[1] https://www.cnbc.com/2025/05/28/zuckerberg-meta-ai-one-billion-monthly-users.html
[2] https://www.theverge.com/news/655930/google-q1-2025-earnings
I am fairly sure that for the google AI usage they count those summaries you get on top of your results are counted in. I doubt it is any different for any other AI vendor. So, just like you suspect highly inflated numbers.
I agree, I usually just scroll the page down quickly on google before the slop actually fills in. They can count it all they want, but it just makes me dislike them more.
To hide the AI slop entirely, I'd recommend either using the uBlock Origin filter list "Huge AI Blocklist" or creating a creating a custom search engine that uses udm14:
https://www.google.com/search?udm=14&q=%s(Or there are also Firefox/Chromium extensions.)
I have considered it, but part of me just wants to make Google lose more money on each search request due to their worthless AI "answer" (yes, I know that this is petty is probably worthless). I do have an extension that filters out known AI slop domains, but I imagine that will get less and less effective over time as LLM usage gets more prolific.
That blocklist just cosmetically hides the AI Overview element; it'll still be generated, if that's your intention.
Excellent, thanks!
Another important note: these massive companies are touting these numbers on earnings calls as if this is fantastic news, but it's good to remember that no one is making money off of all this stuff.
They're not pulling in revenue from end users' AI subscriptions, they're pushing these features in the hopes that at some point they'll figure out how to monetize them. At the same time, the money they're spending on this stuff is absolutely mind boggling to a scale that's literally never been seen before in the technology world. They're all making absolutely massive bets that this technology will somehow make them money soon.
The only company that's actually making serious revenue from AI thus far is Nvidia, because they're the ones selling the shovels in the gold rush.
Either this ends up completely changing the world in really dramatic ways, or there will be a lot of companies that cease to exist in a few years.
I’m betting on both. There were a ton of patently silly VC-burning companies wiped out in the dot com crash, as well as a nontrivial number of potentially decent ones caught up in it too, but the internet as a whole persisted, matured, and changed the world for better and for worse.
Modern machine learning tech is transformative, it’s already winning literal Nobel prizes, but it enables the absolute tsunami of unwanted slop we’re seeing too. I’m not going to predict where we end up because I honestly don’t know, although I will say I’m not at all convinced that chatbots are actually the most important part of the “AI” boom, but either way I do see the ascendancy of neural network based software changing the world just as much as the internet before it and the computer itself before that.
There's some [shitty] middle ground going on, I think. For instance, my Google Workspace account had Gemini shoved down it's throat. I didn't want it, never asked for it, but I sure got it. Along with the price increase per user whether I use it or not. Gemini might fail (and honestly I hope it does, it's terrible), but that per user cost isn't going to go down. Admittedly, that does not offset the full cost of what they've put into Gemini (not even close), but the larger established companies that can afford to burn cash will find ways to mitigate it by forcing it down people's throats.
I don't disagree with your overall premise though. I work for an early stage startup that pivoted to LLM-based products because that's basically the only place VCs are willing to throw money right now and I don't expect it to survive past the next couple years.
Yeah my workplace celebrated us having access to Gemini now so I assume they are paying for it. Now my two sentence emails have two sentence summaries at the top. I'm saving so much time! /s
I’m the Google Workspace admin at my work. We were offered a bundle with Gemini that came out to less per user than we were paying before. That’s the only reason we adopted it. I don’t think anyone in our company actually uses it.
It looks to me like Google is desperate for customers for its AI products. They are effectively paying us to use it.
We are a tech company so we can't stop preaching how amazing AI is. I really hope we got a similar deal. We are big enough it would make sense.
In addition to the crazy amount of money, they're also burning through a crazy amount of electricity. Enough that it's already slowed down the green energy transition. Turning up the heat as the world burns. There's been some reporting on this but I don't think it's a big enough part of the conversation.
Yes, big tech companies all seem to be devoted to pushing unwanted content on people. At best, you get a temporary opt out. (No, Apple, I will never want to turn on Siri.)
I’m wondering where to find better numbers. Maybe look at paid usage?
I use Siri extensively to set timers while cooking and literally nothing else.
If these egotistical companies would just let me name my own fucking assistant I'd adopt one in an instant. But no, every time I turn my lights on I have to say "Okay Google" like some kind of asshole. It sounds like I'm gargling billiard balls.
The branding and recognition of you having to say "Google" as often as possible to remember the brand is worth billions to the company.
Your ick is entirely warranted. It reveals the absolutely nasty intention and aims of the AI companies.
I think too few people feel the ick.
Proven by the fact that this is how we all came to be constantly monetized.
While not totally customizable, Amazon at least allows you to select from a bunch of different Alexa device response options. It makes sense for them to allow that though, as there are sometime several in a house, and it can get pretty annoying when they all respond to the same prompt... especially since (for some weird reason) the further away ones sometimes respond instead of the closest one. Which is why since my parents' responds to the standard 'Alexa', I changed the ones in my area to respond to 'Echo' (which I set to a female Indian accent) and 'Ziggy' (set to a male Australian accent). But you can also have them respond to 'Computer' or 'Amazon' as well.
I have mine set to "Computer," because I love Star Trek. This does have the unfortunate side-effect of it going absolutely berserk whenever I watch Star Trek.
Hey, it may recognize Gurgle as its activation prompt so why not do the billiard balls for real? It'll taste better.
I use it for this and to call people while driving, and that's it, lol. Glad I'm not alone.
I mostly use it for activating my hand-made shortcuts a bit quicker too, but have (very rarely) used it to send a message or a quick call while driving on the freeway so that I don’t have to physically be touching or looking at my phone to quickly give an ETA update when traffic changes things
Meta made the primary message search in whatsapp also an AI message suggestion feature. Whatsapp has checks over a billion users.
It doesn't mean a billion users sought out Meta's AI tools
Is it an issue? Companies are going to brag on their earning calls. I suppose if you're a shareholder?
I was wondering how many users ChatGPT has. It seems this hasn’t been disclosed recently, but Sam Altman hinted at it at a TED talk, reported here.
Since that was six weeks ago, a billion users sometime this summer seems plausible.
Another datapoint:
Anthropic hits $3 billion in annualized revenue on business demand for AI
That's a pretty amazing growth rate.
I don't think there's any public info on profitability, but to compare with money invested, Anthropic raised $3.5 billion in March.
Looks like they've raised something like $12-15 billion since they were founded. (AI-generated table here; I haven't double-checked it.)
The peak of a curve is very hard to predict, and I'm bad at it. (When I joined Google, I thought I had missed the high growth period.) If there's an inherent limit to growth, very high growth will hit the peak faster.
I absolutely love AI, when I mean to use it, I have chatgpt and pi AI apps on my phone
Ive never once used facebooks AI even though they have it on all the posts, and I've seen overviews searching stuff before but if I could turn it off I would
I switched my pixels AI from Gemini or w.e. it's called back to assistant, assistant is so much better, the main reason I use assistant is because when I'm coin roll hunting I can say "1973 Canadian dime" and the first thing it gives me is the link for coins and Canada website which I'm looking for, on Gemini it gives me a blurb I don't need and no option to see links
Plus I like assistants blurbs better than getting geninis blurbs when you ask it a question, I think assistant scrapes it from websites instead of making it up like gemini
The day they get rid of assistant and force gemeni on me is the day I may look into running graphene OS on my pixel out of protest, I'll just Install Google services on it so I can use play store apps, so I won't be going totally free software
I use chat gpt and pi ai every day, but I would like for not everything to be ai, that just makes stuff so annoying for me
Looks like OpenAI's revenue will be at least twice as much as a year ago. (That's assuming they stop growing, but there are no signs of that yet.)
Will they be profitable? It depends how fast they spend.
OpenAI hits $10 billion in annual recurring revenue fueled by ChatGPT growth CNBC
...
So you set everyone's default search to use AI.
Then you can claim a billion AI users.
Truly optimising for metrics above all else.
I know software is less efficient with more resources in modern times, but do I need a bank of RTXs to search for my the café opening hours?
Gotta keep the hype going. There's money to be made.