Hi you,

The exponential developments in A.I. have continued over the past week. To balance out that world, I’ve been walking barefoot outside as much as possible and listening to author and speaker Poonam Sharma’s very human podcast, The Release, which features guest interviews that go well beneath the surface. Here’s what else is grabbing my attention:

• The Washington Post offers the harrowing point of view of doctors in Gaza, and the collapse of its hospital system. It’s important not to look away from such stories. Listening to right-wing voices in the U.S., you’d forget that the foundation of the criticism of Israeli Prime Minister Benjamin Netanyahu’s war is scenes like these. It’s true that there’s been a dramatic rise in antisemitism (well-explored in this NPR conversation featuring my colleague Julia Ioffe). It’s also true that massive crowds in Israel are protesting Netanyahu and demanding a ceasefire and hostage return because they recognize that the future of Israelis and Palestinians is intertwined.
• My friend and extraordinary climate warrior Dr. Ayana Elizabeth Johnson was recently featured in the NYT’s The Interview series. At one point she flips the interview on the interviewer, David Marchese, and asks him, “What are you so afraid of giving up?” when it comes to doing what’s needed to keep Earth livable. It’s a question we should all answer in the face of accelerating climate disaster and the zealous moves of the fossil fuel industry to protect its investments by criminalizing protest. I’ve previously mentioned Johnson’s extraordinary anthology book, All We Can Save, and in a few months, her own book comes out: What If We Get It Right? Visions of Climate Futures.
• While on the subject of women leading in the climate fight, Colette Pichon Battle’s talk at the Bioneers conference is now online. I mentioned her and this extraordinary conference a few weeks ago. Battle is co-founder of Taproot Earth, an organization born out of Hurricane Katrina’s devastation to support frontline community responses to the climate crisis.
• Finally, from the Department of Be Careful What You Wish For, NBC News has the story of the Satanic Temple using the conservative push to insert religion in schools to put Satanism in schools as well. Hilarious. Appropriate. Let’s go. And, if you know, you know. “Bleach Blond, Bad Built, Butch Body.” Thank you, Willonius Hatcher for memorializing this moment between Congressional Representatives Marjorie Taylor Greene and Jasmine Crockett.

His question took on new urgency this week. At Google’s I/O developer conference, on Tuesday, Alphabet announced that it’s effectively going full throttle on artificial intelligence—expanding its own A.I. search summaries to all U.S.-based users before rolling out the function to billions more globally by the end of the year. It was another leap toward what The Verge editor-in-chief Nilay Patel has ominously called “Google Zero”—the day when websites get no more referral traffic from the search giant. And because big A.I. developments come in waves, earlier this week Sam Altman’s OpenAI released GPT-4o, which is essentially a working version of the A.I. assistant from the movie Her. This new version also integrates web searches, so it’s only a matter of time before OpenAI fully takes on web search as well.

Sure, these developments are making headlines. But lately, I’ve been much more curious about the impacts of Meta’s own plunge into A.I. About a month ago, Meta integrated its Meta AI chatbot—which feeds on its large language model, the open-source Llama 3—across its platforms: WhatsApp, Instagram, Facebook, and Messenger. At the top of my WhatsApp chat’s screen, for example, the search function has been replaced by a prompt to “Ask Meta AI or search.” For now, the A.I. responses feel like the basic chatbot stuff of mid- to late 2023. (I can’t believe I’m saying this, but it already feels old.) But it’s wrong to dismiss the relative simplicity of the product today, because as we’ve seen, everything changes all the time in this field—and fast. The bigger question is where this is going, what it means for the web and consumers, and the implications for the increasingly crowded generative A.I. field.

The maneuver also allows C.E.O. Mark Zuckerberg, whose company is valued at just over $1.2 trillion, to play the unusual role of the underdog, fighting for the little guys. As he said in a podcast interview with Morning Brew, in February, his open-source decision is designed to prevent “one organization from getting way more advanced and powerful than everyone else.”

Altruism is obviously not driving this decision. While Zuckerberg has suggested it will take years for the company’s investments in A.I. to turn into profits, Meta’s fortress balance sheet means it has enough runway to see this through. It’s yet another example of the disruptive genius Zuck has demonstrated many times over his company’s 20-year lifespan. (B.T.W., Happy 20th, Facebook! You can almost drink legally.)

Meta is also making a play to radically upend the competitive landscape. By placing its A.I. chatbot in the search field across its apps, it has essentially created a search engine without having to actually build one. Users will now have the ability to experiment with novelty generative A.I. experiences, and to access whatever slurped-up web content is sitting in the Llama 3 training data. In effect, Meta is bringing a new version of the web inside its walled garden, and giving users one less reason to ever leave a place where we already spend way too much time. If consumers take to it, Meta platforms could become users’ primary gateway to the web’s vast ecosystem—a shortcut to information far beyond what’s strictly on Meta’s platforms. And if things really work out for the company, its open-source, price-warring bet will lure engineering talent and business customers to build on its platform, positioning it to figure out where it can take a cut of the value, whether through advertising or some other exchange.

So far, Zuckerberg has decided to wedge this chatbot into the interface and made it practically impossible to remove, a decision amusingly explored by a Detroit local news report. But he’s also taking a big risk. When Snapchat pinned its “My AI” chatbot atop the screens of its hundreds of millions of users, the bot was vulnerable to prurient exploitation that the company worked to fix. Meta is rapidly rolling out a similar product to billions, increasing the risk of spreading misinformation, disinformation, and harm. Because large language models are largely built via unauthorized scraping of other people’s (and companies’) content, there could be massive legal exposure as courts, legislatures, and lawyers hash out the new intellectual property rules.

Notably, Meta is now drifting away from its initial focus on using technology to connect people. There have always been bots on social platforms, but for most of its history, Meta has theoretically been trying to limit those inauthentic communications. Just a few years ago, the company reported that it had removed over a billion fake accounts in a single quarter. Now it’s driving its billions of users to interact with its own non-human account. If a social network prioritizes users’ connections to its A.I. chatbot over their connections with other people, is it still a social network? Or is this just the next step in the humanization of synthetic people, warming us up to a colder world, in which we really are friends with machines?

I don’t know where this is taking us, but I’m confident this signals the end of the Web as we know it. We’re on the brink of a new internet we don’t understand, one in which bots and A.I. agents roam the datasphere on our behalf, synthesizing increasingly synthetically generated content. This “conversation” between Reid Hoffman and his A.I. twin offers a taste of that seemingly dystopian future.

I’m hopeful that we’ll use this moment to discover new ways to credit and compensate creators, whose work has been swept up and spit out by LLMs, and offer us all more control over our own representation; that we’ll move past this trend of companies hurling A.I. spaghetti against the wall; and that our human need to create and connect will be met with a sustainable and reciprocal system that doesn’t just reward the most capitalized players. Meanwhile, we have to stop thinking of the web page as the final destination for information—a reorientation that will either collapse the internet or lead to remarkable and useful innovation. For the latter to come to pass, we’ll need to reorient ourselves away from being simple information consumers toward being active participants in these systems (my whole citizen-is-a-verb thing applies here).

If all we’re trying to do is engage in transactions to efficiently access information, we’ll let the A.I.s run us off a cliff as the companies powering them chase short term profits. If instead we want to be part of thriving communities in which we have relationships with each other (and yes, even brands), then we’ll have to design something more thoughtful than slapping chatbots on everything.