Hi,
I submitted a barrel-filling problem (with a leaking barrel) to Chat GPT and he could succesfully estimate the time necessary to fill it. As this goes beyond pure language processing and demonstrates a certain degree of “reasoning”, how could Chat GPT accomplish this task ?
The latest LLM models (not just ChatGPT but DeepSeek and others as well) have been trained to do advanced reasoning. This is the next frontier in AI. I have actually been experimenting with doing some advanced (PhD level) mathematical proofs in statistics with ChatGPT, and shocked at how good it is at this
1 Like
The language models are now being trained to imitate the research process.
1 Like