You Can’t Tell When AI Is Right
You Can’t Tell When AI Is Right
A summary of an article that never existed, an apology for a mistake that wasn’t one, and why both sounded equally convincing. Some AI answers read as though […]
Read article →ARTICLES
Questions about AI where the interesting part is rarely just technical.
The articles usually start with whatever has caught my attention: a question I keep coming back to, something being debated in the media, or a topic that seems to be getting more attention than depth. DerDenker gives me a reason to look past the headline and work out what we actually know, what is still uncertain, and which part of the story may be missing.
DerDenker is a personal project, not a publication with a content calendar to fill. There is no need to chase every AI story, optimise for clicks or turn every new development into a hot take. I write when something catches my attention — and when I find the time.
Articles → thinking through the implications.
You Can’t Tell When AI Is Right
What Are We Actually Afraid Of?
What Are We Actually Afraid Of?
“AI scares me.” A simple statement that can conceal very different concerns. A closer look at the fears shaping our debate about artificial intelligence.
Read article →
AI Doesn’t Have to Hate UsAI Doesn’t Have to Hate Us
Humans build AI systems, assign tasks and oversee their work. But what happens as those systems become increasingly autonomous? Researchers are already testing how reliable that oversight really is.
Read article →
The Genius That Cannot Count to FifteenThe Genius That Cannot Count to Fifteen
Analyse a complex contract? Often surprisingly well. Reduce a text to exactly fifteen words? Not always so easy. A curious contradiction in the abilities of artificial intelligence.
Read article →
When AI Answers, the Thinking Has Only Just BegunWhen AI Answers, the Thinking Has Only Just Begun
Ask a question, get an answer, move on. Working with AI could be that simple. But there's another way to approach what it tells us.
Read article →
The Data Center as Confessional: Why We Tell AI Things We Hide from Other PeopleThe Data Center as Confessional: Why We Tell AI Things We Hide from Other People
Some things we wouldn't even tell close friends. Yet we tell an AI. What makes a conversation with a machine so different?
Read article →
Cooperation or Flattery? When AI Gets Too Nice to Be UsefulCooperation or Flattery? When AI Gets Too Nice to Be Useful
One mediocre cover letter, three different questions for AI. The assessments vary surprisingly widely. Where does helpful cooperation end and flattery begin?
Read article →
My AI Likes Trump. Yours Might Too.
My AI Likes Trump. Yours Might Too.
Six user profiles, five political questions, 90 responses. ChatGPT was repeatedly asked about Donald Trump's first presidency. Only the information about the person asking changed.
Read article →