Recently, I had the opportunity to glimpse into Altman’s vision from a vast venue located at the northern outskirts of San Francisco. He was present not as the CEO of OpenAI, but as the co-founder and chairman of Tools for Humanity, the umbrella company for a cryptocurrency named Worldcoin, now launched in the US. Worldcoin is available to individuals who undergo an eye scan at the startup’s orbs, which are being introduced in various retail outlets across the country. Simultaneously, the token is pivotal in developing a comprehensive app intended to compete with X.

The World app by Tools for Humanity serves as a crypto wallet, social network, and a platform for mini-applications, all integrated into one. Over the past six months, its monthly user count has soared to 26 million, with 12 million users authenticated through an orb. Within the World app, users can communicate through a chat linked to their World IDs, which is a blockchain-based identity received upon eye scanning, and facilitate cryptocurrency transactions. Additionally, there are countless mini apps provided by developers like Kalshi, enabling in-app transactions using Worldcoin.

Creating a hub for finance and social networking is Musk’s stated intention for X, which is gearing up to launch Venmo-like money features later this year in partnership with Visa. This week, Tools for Humanity also announced a partnership with Visa to release a US debit card later this summer. It will let you transact with Worldcoin and offer rewards for AI subscriptions.

By verifying that people are human, Tools for Humanity is also trying to address the bot problem that Musk said he wanted to fix when he bought Twitter. Onstage at the event, Altman recalled a walk co-founder Alex Blania and he took through San Francisco five years ago, well before the launch of ChatGPT.

“It was clear to us that there was a need for something like this, that we needed some sort of way for authenticating humans in the age of AGI,” he said. “We wanted a way to make sure that humans stayed special and central in a world where the internet was going to have lots of AI-driven content.”

During a press Q&A after the keynote, Tools for Humanity CEO Alex Blania called out X as the kind of platform he’s trying to improve with World ID verification, which the startup plans to eventually charge developers for. Bots on X used to be “so stupid that you could see it was some crypto scam,” he said, acknowledging that now, it’s “not as clear.”

There were several OpenAI leaders in the audience at this week’s event. A rumor leading up to it was that the two companies would announce a partnership for OpenAI’s planned social network. Given the Altman connection, I asked Blania during the press Q&A if he wanted to work with OpenAI. He responded that he was “definitely open to it” and hinted that there would be more to come on that front. I had hoped to ask Altman about it as well, but he slipped out after the keynote.

Earlier this week, I attended Meta’s first-ever AI developer event, LlamaCon, at its headquarters in Menlo Park. That same morning, the standalone Meta AI app was released with a social feed for sharing interactions with the assistant. Eventually, the app is going to add a premium tier and ads.

After his opening keynote at LlamaCon, I caught up with Meta’s chief product officer, Chris Cox, to ask why Meta AI is getting its own app, why it has a social component, and what else he thinks sets the app apart.

The following conversation has been edited for length and clarity:

What’s the goal of doing a standalone Meta AI app?

Almost a billion people are using Meta AI. There’s a bunch of lighter-weight usage, but then there are some communities of people that are going super deep on it. We wanted to offer the lightweight mode in our apps and also the heavyweight mode.

Does this mean that Meta AI’s hooks in the other apps (Instagram, Facebook, WhatsApp) will disappear or become optional?

Our plan is to keep it the way it is. Some of these things make sense in the apps. A lot of people are using it in search to learn more about a specific meme or a specific piece of content. We’ll continue to have those sorts of integrations.

When you compare Meta AI to other assistants like ChatGPT, what do you think is its edge?

There’s voice, there’s personalization, the social features, and then there’s the integration with hardware. We’ve done a lot of work on getting the voice to really low latency and making it really expressive, which lets you have an ongoing conversation.

The second piece is the personalization. It remembers everything I’ve told it. It also incorporates your Facebook and Instagram data, if you connect your accounts, to understand interests in a niche way. It knows I’m a marathoner. It knows I’m very interested in AI. It knows the humor I like. That starts to get you a pretty fascinating level of depth when it comes to answering basic questions. We’ve created the best personalized experiences in social media, and we know that’s the difference between a vanilla app and something that people really love. So that’s going to be a big differentiator, I expect, for us.

The social stuff is incredibly fun. Many people post the equivalent of political cartoons with visual commentary on today’s events. An observation we had is that people don’t know what to do with AI, and it’s not until they’ve been walked through it by somebody who’s an expert that the light bulb goes off. Just like social media, a lot of this is memetic. You’re inspired by what you see other people do.

Will the Meta AI Discover feed become more of its own social network and not just a way to find prompts?

I think we’ll add more to it. I’ve already found a bunch of AI creators using it. With early communities, you want a dedicated place because it’s less pressurized. I’m not going to share my AI-generated artwork on Instagram unless it’s really good. If you have a place that’s just for this thing, there’s going to be a community around it.

Is there hope that this app also helps drive Meta Ray-Ban sales, since it’s how you pair with them?

I think it could in the following way: you see people using Ray-Bans in the app. For a lot of people, that’s going to be the way they get excited about it instead of seeing marketing.

What did you all learn from the Llama 4 benchmark fiasco?

I think the most important thing for benchmarks is to stay focused on what actually matters for your product. Part of what we need to do is just be really, really, really clear on that. I think every organization is going through some version of this. The eval that matters is whether the product is good, period.

If you haven’t already, don’t forget to subscribe to The Verge, which includes unlimited access to Command Line and all of our reporting.

As always, I welcome your feedback, especially if you have thoughts on this issue or a story idea to share. You can respond here or ping me securely on Signal.