Diderot and Smith enter a generative artificial intelligence lounge

Among the many benefits that generative artificial intelligence could bring are the possibilities, available on everyone's phone, of achieving omniscience (which some authors wrongly postulated previously for the internet) and (intellectual) omnipotence, in the sense of allowing any kind of intellectual analysis and synthesis of existing information (obviously, it cannot do analyses that require unavailable equipment; also, it cannot overcome random processes, as a recent study on sports betting has illustrated). It is also becoming omnipresent, being used for all kinds of tasks, all over the world. This combination has raised concerns regarding the impacts on expertise ("The Tragedy of the Cognitive Commons"). It will be pertinent to highlight for the interested reader that there are other "ternary" discussions around artificial intelligence, with titles alluding to three pillars, sometimes with perspectives similar to those considered here, indicating, as key pillars, universality, problem-solving (in K. R. Popper's sense) and explanatory knowledge, or in terms of Predictive, Generative, and Agentic intelligence.  The names Diderot (Denis) and Smith (Adam) are being used here as (hopefully ludic) archetypes of these potentialities, representing information collection ("Encyclopédie", comprising "... des sciences, des arts et des métiers") and performance of a diversity of tasks ("wealth" resulting from "nation of shopkeepers"; the issue of artificial intelligence becoming an "invisible hand" ruling the world, while certainly relevant, will not be considered here, although recent maraudings activities by AI agents certainly highlight it). These theoretical issues are discussed in this work with actual experience, based on empirical results from diverse generative artificial intelligence applications, prompted about questions and tasks in the domain of Earth Sciences, namely in Mineralogy, Petrology (gathering of information in relation to minerals and rocks; analysis and synthesis, including hypothesis of identification, virtual models showing properties and ways of graphical representation ) and Geological Resources (data collection and resource evaluation).  In addition are included some more speculative discussions on CO₂ removal to fight climate change.  This discussion could be pertinent for assessing the potentialities and limitations of generative artificial intelligence as a tool but also for comparison with human minds (possibly suggesting inroads in the study of the latter).