Perhaps LLMs are not as effective at reasoning as appears on the surface due to contamination of evaluation benchmarks (sort of like getting test questions and answers prior to the exam): HERE
Want to use GPT to schedule tasks? Here’s how: HERE
AI can make that problem “disappear” for you.. : HERE
OpenAI wishes to box-out the competition via regulation: HERE
Zuckerberg’s comments on AI replacing/augmenting mid-level engineers in 2025 (below):