An AI in Every Pot
/cdn.vox-cdn.com/uploads/chorus_asset/file/24987916/Mark_Zuckerberg_Meta_AI_assistant.jpg)
Another day, another new model. In this case, Meta's "Llama 3". But underlying technology aside, this strikes me as a big deal because of how Meta is bringing it forward. That is: basically everywhere they can.
Even though it just launched yesterday, I already have encountered the new AI in both Messenger and WhatsApp. It's not yet in my Instagram, but it will be soon. Same with Threads. Oh yes, and Facebook itself. Meta even now has a website to play around with the technology directly. EVERYWHERE.
That matters because the company, through its services, has a scale that not even Apple can match. Billions of users will get this AI shoved in their faces when they roll it out worldwide.1 And this matters because it will give Meta a scale advantage when it comes to data.
And Meta remains, at its core, a consumer company. What strikes me most about these takes on AI is how simple and well designed they are for regular usage. A lot of the first wave of AI tools have been comically obtuse. I believe that's a big part of what helped OpenAI surge out of the gate – first DALL-E and then ChatGPT were just simple, intuitive consumer products.2 Now Meta is putting a coat of paint on such tools just as they would with Messenger and Instagram. The result is AI that looks inviting, almost playful. That matters, obviously.
Other tidbits from Alex Heath:
The Meta AI assistant is the only chatbot I know of that now integrates real-time search results from both Bing and Google — Meta decides when either search engine is used to answer a prompt. Its image generation has also been upgraded to create animations (essentially GIFs), and high-res images now generate on the fly as you type. Meanwhile, a Perplexity-inspired panel of prompt suggestions when you first open a chat window is meant to “demystify what a general-purpose chatbot can do,” says Meta’s head of generative AI, Ahmad Al-Dahle.
Again, all things that make this decidedly usable and useful.
There’s a comparison to be made here to Stories and Reels, two era-defining social media formats that were both pioneered by upstarts — Snapchat and TikTok, respectively — and then tacked onto Meta’s apps in a way that made them even more ubiquitous.
Some would call this shameless copying. But it’s clear that Zuckerberg sees Meta’s vast scale, coupled with its ability to quickly adapt to new trends, as its competitive edge. And he’s following that same playbook with Meta AI by putting it everywhere and investing aggressively in foundational models.
“I don’t think that today many people really think about Meta AI when they think about the main AI assistants that people use,” he admits. “But I think that this is the moment where we’re really going to start introducing it to a lot of people, and I expect it to be quite a major product.”
It's worth noting that Meta has tried to roll out various element of AI previously – even dating back nearly a decade ago, well before the ChatGPT revolution, of course, with "M". More recently, their celebrity AI bots have failed to take the world by storm. But this new push feels more correct. Especially as the LLM arms race keeps pushing things forward while pushing the technology further into the background. Again, the actual front-end is what is going to matter to 99.9% of people here. And Meta both knows this and knows how to do this.
The pace of change with AI models is moving so fast that, even if Meta is reasserting itself atop the open-source leaderboard with Llama 3 for now, who knows what tomorrow brings. OpenAI is rumored to be readying GPT-5, which could leapfrog the rest of the industry again. When I ask Zuckerberg about this, he says Meta is already thinking about Llama 4 and 5. To him, it’s a marathon and not a sprint.
“At this point, our goal is not to compete with the open source models,” he says. “It’s to compete with everything out there and to be the leading AI in the world.”
To me, one interesting next question in all of this is if Meta can take advantage of their software scale to bolster their own AI, can Apple do the same on the hardware side? We'll perhaps get our first glimpse at that answer in a few weeks at WWDC. But beyond AI-generated wallpapers and the like, what I really want to know about is how Apple is going to use the iPhone (and other devices) out in the physical world to help train their own systems. Again, this is something Apple can uniquely do at scale.3 And it's the flipside of what all these AI companies want to do more with hardware – including Meta, of course. How much will this matter?
1 Interestingly, even here in the UK, I have to VPN in to use it still.
2 But perhaps some challenges lay ahead...
3 Clearly, Google will be trying as well with their Pixel devices, but those aren't at the same scale. Samsung is perhaps closer, but seemingly relying on Google quite a bit for their AI. Microsoft might make an interesting play by leveraging Windows, but it's Windows.