Teach your children to click links

The highlight of the week? Our first New Friends dinner, where the conversation reminded me just how much I love the web. And you do too—you just don’t think about it much. More on that below!

Then, Apple’s upcoming home “experience,” a way to confirm AI-generated images and an Old Thing with a surprisingly fun backstory.

🎶 “I believe children are the future, teach them how to click links and let them lead the way…” 🎶

Those are definitely the real lyrics to Whitney Houston’s classic.

This week, I found myself thinking a lot about how we need to teach the next generation to visit webpages. The future of the web may depend on it.

That was the theme of our first New Things’ New Friends dinner: the future of the open web in an AI world. The dinner sponsor, Yahoo, was particularly appropriate. Its Scout AI answer engine is designed to surface publishers and drive people back to the original sources.

The problem is, the future is looking less and less like link-clicking. Take my own use of Meta’s Muse and OpenAI’s Dot. I have both summarizing the top tech stories for me every day. Muse’s Feed tab even serves up a ready-made summary of things it thinks I’m interested in. 

And even if you aren’t using cutting-edge agents, chances are you’re using Google’s AI Overviews or another chatbot to answer questions—and visiting primary sources less often.

The dinner conversation was held under Chatham House Rule, which means I can tell you what was said, just not who said it. The room included senior folks from major tech companies, along with leaders in venture capital and media.

Here’s why I’m worried about the web—and why I still have some hope.

The fear 

Traffic to websites is dropping faster than my kids’ backpacks in my kitchen after school. For consumers, AI answer engines mean convenience and time saved. For web publishers, it means fewer human visitors—and fewer ways to make money. My friend Nilay Patel calls this Google Zero. Google will eventually send ZERO traffic back to the websites that supply its answers.

Now, you might argue: OK, but if my AI agent visits the website and gets the information for me, isn't that still a visit? Technically, sure. But here's the problem: Your AI agent doesn't care about the ads on that website. It's grabbing what it needs and getting the heck out. And advertising, love it or hate it, has helped keep much of the internet free for decades. 

If the future of computing involves fewer screens and more AI voices whispering information into our ears, glasses or whatever new gadget they're dreaming up, there could someday be no links to click on, no ads to see. And potentially no easy way to know where that information came from in the first place—not great when trust in online info has already been sliding for decades.

The hope 

Someone at the dinner made a point that stuck with me: Normal people—and by that I mean people who don’t spend their days figuring out how to get AI agents to do everything short of wiping their butts—aren’t sitting around thinking about how much they love the open web.

But here’s what people do love: browsing. Window shopping. Scrolling. Sometimes we don’t want an AI agent to find the perfect pair of shoes and buy them in three seconds. We want to scroll through 47 pairs we’ll never buy.

Someone even brought up a website called “Food Never Comes.” It lets you browse fake menus, place fake orders with fake credit cards and even track a fake delivery driver. The food, as it says, never comes. The whole point is to get the dopamine hit of ordering takeout—without spending a dime. Or consuming a calorie.

Some product leaders said they’re deliberately making their AI summaries shorter, and even boring, to encourage people to visit the original source. And there was big optimism for journalists, podcasters and creators. An AI agent can summarize the news, but it can’t necessarily replace the personality, perspective and trust that bring people back to a particular human. Someone invoked Maslow’s hierarchy of needs: If content delivers entertainment, connection or fulfillment—not just information—people will keep seeking out the people who create it. That's encouraging news for me!

There’s also a reality check amid all this AI hype: Fewer than 5% of Americans pay for even one AI service, according to a new report from Andreessen Horowitz. For all the talk of AI taking over our lives, most people aren’t even willing to pay for it yet.

Still, let’s please teach our kids to click links. Will that alone save the open web? Probably not. But hey, at least they’ll know what the web looks like—and how to dodge 87 cookie pop-ups just to read one article. 

Apple sent out invites Thursday, including to yours truly, for an “experience”—don’t you dare call it an “event”! It is this coming Tuesday, October 13 in New York City. The title? “Welcome home.” The invite featured a fun hand-stitched Apple logo—or what I assume is officially called an iPatch.

There have been reports, mostly from Bloomberg’s GOAT Marc Gurman, that Apple has been cooking up a smart-home hub with a touchscreen—possibly, and hopefully, called the HomePad. Remember that one from our September Apple event drinking game? Also, expect a new Apple TV. 

I’ll be experiencing the experience (experientially, of course) and bringing you all the news in Wednesday’s newsletter.

See if media is AI-generated

Earlier this week Google made its AI detection tool SynthID public for all users. Developed by Google DeepMind, it embeds invisible watermarks in AI-generated content, making it possible to identify its origins. Google has been collaborating with other companies on the watermarking standard, so the tool can now detect AI-generated content from not just Google but also OpenAI, Nvidia and others.

To try it out, head to SynthID.com, drag and drop any photo, audio or video file into the “Upload for file detection” box and it will tell you if the content was AI-generated. 

When we put the tool to the test, it identified a dog on the beach in sandals, generated with ChatGPT, within seconds. The New Things logo? SynthID correctly said no, since all of our branding was done by the brilliant humans at Brainstorm.

Attendees of our first New Friends dinner all were told to bring Old Things. Here’s one from a leading tech founder and engineer, with a gadget backstory I particularly loved.

Name: David Singleton 

What's your old thing? Sharp Pocket Computer PC-1425 

What year is your old thing from? 1986 

Why do you love your old thing? This is how I learned to program. My dad was interested in buying one of those big over-the-shoulder VHS camcorders, and he liked to strike a deal when he was buying something. So he tried to get money off this camcorder, and the store said, “We can’t do that, but you know what? We have this thing that nobody wants. It’s a programmable calculator.”

He agreed because that made the camcorder more palatable, and so he gave the calculator to me. It had this remarkable feature where you program it in BASIC, but when you press the keys, they each expand to a BASIC keyword. So you press P for print, and it expands. And then you can learn what the parameters are by trial and error. I made little simple programs with this thing. Months go by, and I show my dad a really cool computer game I built on it. He’s like, “Oh, do you want the manual?” This product was very formative for me.

Our gift to you, David: a PDF of the Sharp PC-1425 manual.

This newsletter was written and curated by Joanna Stern and Adele Lowitz.