June 1, 2026
Brains, bugs, and bruised egos
The Frame Problem
AI says this old brain teaser is solved, but the comments are absolutely not
TLDR: The article revisits an old AI problem about telling what changes and what stays the same after an action, a puzzle that later spilled into bigger questions about how humans think. In the comments, people split between “this was solved ages ago” and “not so fast,” with one killer joke suggesting humans aren’t exactly masters of relevance either.
This wasn’t just a dusty philosophy-and-computer-science explainer — it turned into a mini custody battle over an old intellectual headache. The article walks through the so-called “frame problem,” which is basically the question of how a machine can know what changes after an action without tediously listing everything that doesn’t change. In plain English: if you paint a thing blue and then move it to the garden, common sense says it should still be blue. Early AI systems, hilariously, needed to be told that moving something does not secretly repaint it. That awkward gap sparked decades of debate between computer scientists and philosophers over whether this was a technical bug, a deep truth about thought, or both.
And the commenters? Oh, they had thoughts. One of the sharpest jabs came from MarkusQ, who basically framed the whole saga as philosophers standing on the porch yelling, “AI people don’t want this problem anymore, but we’d like it back please.” That set the mood: half academic roast, half breakup drama. Others argued this is old news and newer ways of describing changing situations may have already patched the hole. But not everyone was buying the “humans are great at relevance” premise in the first place — one deadpan reply shot back, “I don’t think we really have such an ability. Hence microplastics.” That line pretty much stole the thread. So yes, the article is about logic, but the real spectacle is the comment section asking whether this famous puzzle was solved, sidestepped, or simply rebranded until everyone got tired and moved on.
Key Points
- •The article defines the frame problem in AI as representing action effects in logic without explicitly listing all obvious non-effects.
- •The problem originated in logic-based artificial intelligence and was later broadened by philosophers into a wider issue about relevance and reasoning.
- •The article says debate between AI researchers and philosophers was especially active in the 1980s and 1990s.
- •An example using painting and moving an object shows that classical predicate logic does not automatically preserve unchanged properties such as color after movement.
- •Frame axioms can encode non-effects explicitly, but the article argues this approach scales poorly because the number of required axioms grows with the number of actions and properties.