A researcher in chemistry shared their experience using free AI tools to solve complex scientific problems. By engaging in iterative dialogue with the model, the user successfully derived a new scientific concept that explains phenomena previously unresolved by existing theories. The proposed mechanism aligns with both published literature and unpublished experimental data.
In a separate case, the researcher utilized AI to develop a new mathematical equation for temperature dependence. The goal was to combine high precision with extrapolation capabilities, a feat that historical models by Einstein and Debye could not achieve. The AI generated a solution that met all specified parameters, including constraints on the number of variables and derivative behavior.
The author emphasizes that these results were achieved using free, publicly available AI interfaces without feeding the models specific datasets. While the AI models frequently hallucinated scientific citations, the underlying logic and creative contributions were sufficient to qualify as significant scientific work. The author suggests that AI's ability to generate such solutions challenges current assumptions about its limitations.
