I'm happy for you tho, or sorry that happened
To answer seriously, it's awful, absolutely awful, to a level which could only be achieved by AI mimics.
I suggest you take the time to seriously learn how LLM's work. Know what you are working with. They are probability engines, and have no way to determine what happens in "2066/2067", etc. Look, I spend my days trading financial markets, and work with LLMs... they can't predict the next day of human events. They generalize based on backwards looking knowledge.
Humans are forward looking. If you want to empower yourself, stop using backwards looking tools literally using baked in probabilistic responses based on historical data, and start pushing forward into new territory.
You can lay out tests easily. Ask what will happen with the Modi/US tariff negotiations, and the Putin/Ukraine war, and ask for specific dates. You will soon see them violated. Or better yet, review the code. Literally look at the code and see how the probability machine works. It's not magic. You're basically like someone worshipping Google in year 2001.
You would be much better off reading legitimate philosophical texts if you want to gain enlightenment. You are giving up your autonomy and humanness by subjugating yourself to the sum total of calcified historical inputs. Live in the moment, and push forwards. Stagnation is death. In many ways, I see this post as a manifestation of true "evil". That's my view, since you asked.
That said, it's feasible that an advanced system (not LLMs probably) could make accurate long term predictions and inaccurate short term predictions.
In Asimov's psychohistory concept, the predictions got more accurate the further into the future they looked, because individual human actions are unpredictable, but the aggregate behavior of billions of people follows statistical patterns that become more reliable over longer timeframes. Short term predictions are vulnerable to random people/events that average out over enough time.
And when predicting the distant future we go broad (social trends, foundational technologies, etc) but when predicting the immediate future we go narrow (what'll NVDA be worth next week).
I’d love to know what you hear in it.