Tech people and their friends are like 1% of the population.
Go out on the the street of Anytown in any western country, and people know "ChatGPT".
A friend of mine is a teacher, and told me that at a recent school board meeting there was discussion about implementing AI into the learning curriculum. And to the board, "AI" and "ChatGPT" were used interchangeably. There was no discussion of other providers or models, because "AI" is "ChatGPT".
That's why OpenAI has these huge projections. When average people are asked to reach for AI, they reach for ChatGPT.
> When average people are asked to reach for AI, they reach for ChatGPT.
No, average people are nowhere near that tech-savvy. Just because every mom in the 90s called every video game console a "Nintendo" did not mean that Sony didn't mop the floor with Nintendo in that era. This isn't brand loyalty, it's brand genericity. Other than, say, Replika-style users who have formed an emotional bond with a certain style of chatbot, no average joe on the planet gives a damn whether the LLM powering their chat is provided by OpenAI or Google or etc. They'll use whatever's in front of them and most convenient, and unlike Google, Apple, or Microsoft, OpenAI doesn't own the platform that establishes the crucial defaults that nearly no user ever changes.
Except Open AI happen to be the Sony in this case. 700M Weekly active users and the 5th most visited site on the planet, with no-one else close. I mean, it's pretty clear this is less 'Nintendo' and more 'Google'.
This isn't a rebuttal to any of the above criticisms. You appear to have completely misunderstood the existential risks that OpenAI faces here.
Well you call this instance brand genericity instead of brand loyalty and hinge your argument on a past situation that doesn't really have any resemblance to this one.
Gemini is a google default so why isn't it used anywhere near a much as chatGPT ?
Meta has stuffed their llama model into Instagram, Whatsapp, Facebook and god knows what so why isn't it used anywhere near as much as chatGPT ?
In all the time these players pushed their apps and models to their billions of users, chatGPT's userbase has been growing massively.
Clearly people do care about using chatGPT and specifically chatGPT and what would be the 'existential crisis' by the incumbent players has come and gone, with Open AI unscathed.
If Google was massively unprofitable and with no profitability path in sight, yeah, that's very much Google.
Google was unprofitable for some time yes. I don't understand the obsession with the 'no profitability path in sight'. It's nonsensical. Their reported loss numbers are really low for a service with that many users. Open AI don't have to do a lot to be profitable, ads would be enough.
> Open AI don't have to do a lot to be profitable, ads would be enough.
Yeah man, absolutely! Ads can pay for all of it!
> It's nonsensical
Yep, wanting a business to be profitable before investing in it, or at least to show that it could be profitable by providing a plan and timeline to profitability, is too much to ask for nowadays. Of course companies can be successful and great without any clear path to profitability, hype and enthusiasm are enough, just look at the great success of WeWork and Adam Neumann!
Who are we to doubt that? Just keep repeatedly giving couple of trillions to Sama and all will be okay, AGI is just behind the corner, trust him bro.
OpenAI reportedly made a loss of $5B in 2024. By 2024's end they had ~300M WAUs, most of them free and not monetized in any way whatsoever.
To be profitable, they'd need to monetize those free users for an average of $17/year or $1.5/month.
Do you realize how low that is ? Yes, ads even implemented in an uninspiring fashion would more than cover that. With some thought into it ? Well, You don't need to be a genius to see the potential of directly weaving paid recommendations naturally into conversations when appropriate.
Exactly, it's hard to dismiss the broad penetration of ChatGPT in the general population. I was an AI skeptic/luddite until almost exactly a year ago when, in a span of a month or so, I had three different friends/family members who work in various administrative jobs tell me that they all used ChatGPT surreptitiously at work to get things done. Now a year later I don't know many people that don't use it at least occasionally. The ones that don't are older and I'm confident eventually they'll be using it like crazy to annoy me.
Anecdotally, "skype" was once synonymous with video calls but it's pretty much never used now.
It's literally never used now, because it was taken offline earlier this year.
And before that it was essentially irrelevant and on life support for what, maybe a decade?
> because "AI" is "ChatGPT".
People still call it "Kleenex" when they're using any old facial tissue. They may still call it "ChatGPT" when it's coming from Google.
They may be satisfied with anyone’s “ChatGPT” though.
More than just tech people on X and they all know Grok.
I imagine Meta users know Llama too?
Facebook users do not know Llama anywhere near the degree the X network does Grok, and both are pallid in comparison to Chatgpt.