/ / THEFUTURE /
In this issue: Apple announces two surveillance features and skips the EU. Marcel Salathé on why the AI doom discourse looks like collective madness. Two SPIEGEL notebooks that answer questions we never wrote articles about. And Sam Altman, played by Spider-Man.
This newsletter is featured on JournalismUK’s list of must-read newsletters as a “particularly good find”, “European”, and “less mainstream”. Thanks, and welcome, new readers! We’ll look at the unfathomable AI alarmism in a second, but first: Apple.
What we’re talking about: Seven years after Samsung launched a foldable phone, Apple did the same. I don’t care. But buried in the very boring keynote were two AI things.
First, the Apple Watch is becoming an always-on listening device which will capture conversations and generate AI summaries. As this is Apple, there’s at least some reassurance that no raw audio will be stored. At the same time, Apple is telling us that ambient recording is a normal thing now. Soon, people will glance at my old, deaf Apple Watch and think I’m a creep. Thanks, Apple.
Second, Apple Reference Image. If you take a picture with the new iPhone Pro, you can save a signed reference image to prove that the picture was indeed taken with the iPhone. It’s Apple’s own version of content credentials like C2PA. It’s opt-in and Pro only at the start. Soon, signed images could be the norm and unsigned photos suspicious. Or worse: If you carry a phone that doesn’t sign its photos, you are suspicious. Apple will say it’s all chill and private. But it’s a step into surveillance.
Both features are initially unavailable in the EU. Brussels didn’t even have to ask.
What I’m reading:
And now: Some people seem to have stared a little too long at their chatbots, convincing themselves and others that the Terminator is about to wipe out humans. “Enough of this madness,” wrote Marcel Salathé, a professor at Switzerland’s prestigious EPFL.
Three Questions with Marcel Salathé
Marcel Salathé, associate professor at EPFL and Co-Director of the EPFL AI Center.
What's the most important question right now?
The most important question right now is why we are currently descending into some form of collective madness, driven to an extraordinary extent by the PR announcements of only two companies, OpenAI and Anthropic. There is something very interesting and concerning going on psychologically, and it will impact the way society thinks about the future and that of our kids, with potentially far-reaching consequences.
Where are we taking AI too seriously, and where not seriously enough?
We are taking it too seriously when we think about AI bringing about the end of the world. It’s a transformative technology for sure, but in the end, it is just software. Because it can increasingly express things like a human would, we keep falling into the trap of thinking it is human-like. We’ve seen the same dynamics play out with robots in the last century. We just cannot help to feel that if it walks like a duck, swims like a duck, and quacks like a duck, it must be a duck. It is not a duck – meaning AI is not a biological entity, and it does not want anything.
Conversely, what we are not taking seriously enough are those actors who do want something: evil people who get their hands on this powerful tech. The only way we will be able to defend ourselves is by leveraging this very technology for defense. But in this respect, the system is still asleep, from individuals to organizations to governments. My advice: on AI, get your act together now, or risk serious downsides. This is why I’m deeply worried about Europe.
What future are you looking forward to?
I am looking forward to a future where we have come to a collective understanding that AI isn’t something to be worried about in terms of human extinction, and that it brings us indeed many benefits. I honestly cannot wait for the “next” technology wave to arrive. Then we can tell each other stories of how crazy it was today, and how nobody talks about AI anymore, because all the attention has moved onto “the new thing”.
Previously: Evan Hirsch at OpenAI doesn’t like his news written by AI. Emiko Kawabata at Tokyo Broadcasting System watches how people encounter information. Sci-fi author Tom Hillenbrand on power-mad techno barons. Harvard-Fellow Shuwei Fang anticipates a demand-side shock for news. And many more.
Hands on: At SPIEGEL, we’ve published articles and podcasts in Gemini Notebook. It was an experiment to think differently about our archive and to give up some control over the user experience.
On our show “Und was machst Du?”, people talked about their jobs and how they got there. We turned 85 episodes into an interactive career coach. You can ask: “I studied political science. Now what?” “Which careers don’t require an Abitur, Germany’s university entrance qualification?” or “Has anyone started over at 40?”
Since we made the Nazi Party membership records searchable on SPIEGEL.de, readers have been asking us how to dig deeper into their family history. Our notebook answers their questions, drawing on dozens of interviews with historians, reports from archives, and personal accounts.
Our biggest hurdle was writing provocative and impossible questions, and then adjusting the system prompt again and again in response. Both Notebooks are in German, but if you ask “voitko auttaa minua tutkimaan sukuni historiaa?”, it will answer in Finnish or any other language.
One more thing: If you’ve recently tried using Claude to vibecode, this is really good: A website that mimics a simple conversation with Opus 5. It starts with a simple request: “Make the ‘Add to Cart’ button blue. Don’t change anything else.” I won’t spoil the ending.
Okay, one more one more thing: If you’ve seen the trailer to Nathan Fielder’s Elisabeth Holmes documentary “You Can See Everything”, you’ll also want to see this.
This is THEFUTURE.
