OpenAI's o3 model might be costlier to run than originally estimated | TechCrunchOpenAI's o3 AI model exhibits costlier computation than initially thought, with potential costs soaring to $30,000 per task.The revised estimates indicate significant resource demands and potential pricing challenges for advanced AI models.
AI researcher Francois Chollet is co-founding a nonprofit to build benchmarks for AGI | TechCrunchFrançois Chollet's ARC Prize Foundation aims to develop benchmarks for assessing AI's approach to human-level intelligence.
A test for AGI is closer to being solved - but it may be flawed | TechCrunchThe ARC-AGI benchmark shows limitations of AI tests, particularly focusing on memorization rather than true reasoning capabilities in language models.