I Stumped All AI Models with My First-Grader's Homework
Sitting at the dining table helping my daughter with homework when I hit a question with multiple possible interpretations. Curious, I tested it on several AI models - they all failed spectacularly.
Tell me what you think? What's the right answer. See below to see how AI fared.
Claude 3.7 Sonnet
OpenAI o1 Pro
OpenAI o3-mini-high
OpenAI 4.5
OpenAI o1
OpenAI 4o
Gemini Flash 2.0
Reduce risk - focus on vulnerabilities that matter - Contextual ASPM - CEO & Founder - Phoenix security - ??♂? Runner - ?? Application Security Cloud Security | 40 under 40 | CSA UK Board | CSCP Podcast Host
1 周Caleb????
Principal Engineer at ARM
1 周This isn't really a 1st grader test. It is an IQ test questions from the looks of it. It does seem very difficult from a spacial reasoning perspective. If you know, you know, if you don't, I believe that AI would struggle with this question.
I build effective cybersecurity programs, exceptional teams, and rational processes
2 周Can we please stop confusing ourselves by testing non-deterministic generative technologies with deterministic tasks? It does not prove that AI is stupid. It proves that we are.
Protecting the technology to electrify & decarbonize the planet @ GE Vernova
2 周“C” - as you rotate the figure you would get all other representations (A,B,D) except for C
Entrepreneur | Investor
2 周Given the question is so broad, I would default to as simple an explanation as possible especially given the context that this was asked of a first-grader (context is king ???? imo) — not that other answers aren’t correct as well. In my interpretation, I would say A is the “correct” answer because all the rest have one square in the second row from the bottom while A has three. Happy to be wrong, would love to hear why?