No it does do calculations. In fact, you can even ask it to show you step by step how it arrived at the answer. Heck, you can even ask it to write a program that does the calculation.
This is how out of touch commentators are on Macrumors about state of AI.
It’s fine to be keen on AI, knock yourself out, but you should educate yourself about how they really work.
They don’t think or calculate anything.
As for your previous post asking me for an example, I can give you one from this morning.
For context I am a software developer, and I am very much aware of the strengths (and imo very big) weaknesses.
I was expanding the functionality of a certain service class which listens to updates from a few different repositories and does certain things with the data, in order to separate the concerns of different parts of the codebase into one point of access.
Now the problem, or better, idea of what was to be built wasn’t exactly trivial, and the solutions offered by the ai were, on paper, not stupid, and technically worked. But they massively over complicated not only the functions themselves, but the responsibilities of the class. We went around in prompt circles trying to fix this problem, as is always the case.
The fix?
Taking my dog for a walk and having the idea of a completely different way to architect the solution, using a separate class entirely.
I would have gone around in circles with the ai for hours and ended up with an abomination of coupling and over complexity, because they can’t, I repeat, cant reason, think, come up with ideas or real solutions.
It heard me say: I want to add this feature/functionality to this context, and it spewed out the appropriate slop for that scenario.
At no point is it possible for it to say - hang on a sec, is this a good idea? Maybe you should put this somewhere else or think about it differently. Because, why? It can’t think. It can’t have ideas.
