From my understanding, most LLMs work by repeatedly putting the processing output back into the input until the result is good enough. This means that in many ways the input and the output are the same thing from the perspective of the LLM and therefore inseparable.
The title of this article just doesn’t match reality. It really only (maybe) applies to very high end systems that are already pushing the limits of all components. Most people don’t have the money to waste on that and have plenty of room to upgrade their hardware for a looong time.
If you don’t need much (e.g. no gaming, 3D rendering, etc.), especially if you don’t need a dedicated gpu, then you can upgrade for at least a decade before running into issues. To be fair, a laptop should last a decade as well in that case, but at a higher prices and while being less repairable.
The previous comment gives a pretty clear argument for why desktops are more future proof, I think. Being more repairable is a pretty big deal for the longevity of the whole system.
Although I agree with the sickening greed part, I don’t think it makes sense to make educational use free without have another system in place that pays for the writing of educational books. There’s plenty of content that imo should be free for educational use, but educational books only have an income from educational use, it’s their whole target audience. No income, no book :/
Ah yes, ‘best technologies in the world’ like the software giving Google and the USA full access to all our data?