Ole Reissmann

About · Newsletter

THEFUTURE

Fakecasts and Rainbow Sparkles

Newsletter sent 22.7.2025 by oler

In this issue: AI Overviews invade Google Discover while AI Mode sparkles for attention. Why some journalists are living in an AI bubble while others remain oblivious. Alexandra Borchardt on the three strategies every newsroom needs. Plus: I created a “fakecast” and immediately regretted it.

What we’re talking about: Google has launched AI Overviews in Discover. It’s kind of a big deal. When someone searches on Google and visits your website, that’s Search. Then there’s a personalized feed in the Google app with suggested articles—that’s Discover.

For some publishers, Discover really is important. Even more so now that AI Overviews are eating up real estate in Search, blending and bundling articles into summaries. Until now, Discover was big pictures and single articles. It was fun while it lasted.

AI Overviews in Discover are currently live in the US, focusing on lifestyle and entertainment, as TechCrunch confirmed with Google.

But another US launch could be even more consequential: AI Mode, Google’s search chatbot.

AI Mode gets promoted right in the search bar and at the top of search result pages. It sparkles in rainbow colors to grab attention and pull you into a chat interface. No more search results, just answers.

Is this the future of search? Or just Google trying really hard to make AI happen? Either way, it’s a whole new Google.

What else I’ve been reading:

AI & Journalism Links

AI is killing the web. Can anything save it? asks The Economist, referring to business models that rely on traffic and advertising. Tollbit, a paywall for bots, reports its highest per-crawl rates at a local newspaper. Unique content appears to be part of the solution.

AI etiquette: “Whoa, let me stop you right here buddy, what you’re doing here is extremely, horribly rude.” (Alex Martsinovich)

The Directory of Liquid Content: Sannuta Raghu made a framework to make it easier to assemble, adapt, or repurpose news stories across contexts, users, or platforms.

How the Associated Press Built its AI Strategy Without Breaking Trust (Ulrike Langer, News Machines)

AI-generated band “The Velvet Sundown” garners 1M+ Spotify streams, sparking debates over platform transparency and artist compensation. (Lanre Bakare, The Guardian)

Strategy over shiny objects: Newsrooms are split between sophisticated AI projects and journalists who have no clue what’s happening down the hall, says Alexandra Borchardt. Her advice? Start with strategy, not tech—and maybe listen to some birds once in a while.

Three Questions with Alexandra Borchardt

Hands on: I did a little experiment, creating a podcast where two AI personas discuss content based on a newsletter and three related articles. I fed a detailed prompt into a large Claude model, then ran the output through ElevenLabs and Gemini text-to-speech. Essentially, AI talking to itself.

You can find the MP3, transcript, and all prompts here.

Midway through, it hit me: I’m cool with human-like voices narrating text, but mimicking an actual conversation? That feels different. Uncanny. It’s not a podcast. It’s not an audio overview. It’s a “fakecast.” (Trademark pending. Just kidding. But seriously, let’s make that a thing.)

Whether this is good or bad is kind of irrelevant. ElevenLabs and Google are actively pushing these features. But still, I can’t help but wonder—will this all feel like a gimmick in a few months? Anyway, if nothing else, let’s at least agree to call them by their name. Say it with me: fakecasts. Catchy, right?

One more thing: You know what word I can’t stop associating with AI lately? “Dropped”—or better yet, “just dropped.” A company launches something. A paper gets published. Doesn’t matter what it is, it gets dropped. Like an Amazon package into your neighbor’s garden.

What’s supposed to sound like “Drop it like it’s hot” now sounds more like carelessly discarded trash to me. Thanks for nothing, AI.

A very good post from the Nuclear Threat Initiative on Threads:

"I asked chatgpt" "I asked grok" well I asked Jared Harris from the critically acclaimed HBO miniseries Chernobyl and he said this is how an RBMK reactor core explodes.

This is THEFUTURE.

Subscribe to THEFUTURE

Get this newsletter in your inbox. For free.

The previous issue is AI’s Next Land Grab: Your Browser, the next issue is Magic Tokens and Modern Luddites.