All Posts 5
Essays
Deep Dives
Explainers
Dialogues
Featured
Can AI Ever Truly Reason? A Conversation at the Edge of Logic
A wide-ranging conversation exploring whether current AI systems are capable of causal reasoning, what counterfactuals are and why they matter, and why the Liar's Paradox broke mathematics, computer science, and may set fundamental limits on artificial intelligence.
More from SquareCircle
Gödel's Incompleteness: The Theorem That Broke Everything
In 1931, a 25-year-old mathematician proved that every sufficiently complex logical system must contain truths it cannot prove. Here's why that matters for AI, consciousness, and the nature of knowledge itself.
The Brilliant Parrot Problem: Scaling vs. Understanding in AI
More parameters, more data, more compute — but are we getting closer to genuine understanding, or just a more convincing performance? The case for and against the scaling hypothesis.
What Are Counterfactuals — And Why Can't AI Think in Them?
The ability to reason about what didn't happen is at the core of human intelligence. Science, law, medicine, and morality all depend on it. Here's why it's so hard to build into machines.
I Could Draw a Square Circle — But I Won't
On the difference between logical impossibility, physical representation, and willful refusal. Why the most interesting answer to an impossible question is sometimes "I could, but I won't."