Context is always important. As an EE, I quite often approximate 1s with infinity e.g. this capacitor will have zero charge as t approaches infinity which I can assume is definitely after 1s. Not sure if all EEs think about it that way, but this actually kind of what we are doing.

Cells don't need to, as they can be… "renewed", you might say.

They don't get to be renewed with zero pollution.

All life processes produce waste that is dangerous to that which created it, yet we do not call photosynthesis "non renewable" despite oxygen being toxic to obligate anaerobes.

Likewise while PV has non-zero impact, the impact is sufficiently small and of a nature that basically doesn't matter.

Fossil fuels are about a hundred times more polluting than volcanoes[0][1], so a factor of about one hundred in reduced emissions per unit of useful output compared to setting things on fire takes you down to the environmental impact of one part of geology.

[0] at least the CO2 is, but what matters for PV is everything else except the CO2 because in a world powered by renewables all the energy used in production doesn't produce CO2 in the first place; volcanoes do also release sulphur compounds etc.

[1] there is a certain degree of irony that one of the memes used by climate change deniers is to precisely invert this ratio.

Well, probably nothing we build will. And probably nothing we say will last forever, actually last forever. We still use the word forever and other unprecise terms in common language. I also hate it at times, how unprecise (and wrong) normal language is. But I have given up on the habit of getting mad about it and rather go with the intent and not about the single words.

So no one meant literally indefinite energy here, it was meant, "potentially way more energy, than we currently need". And this is correct.