Back in the day, I had a teacher using spreadsheet software for a grade book that could automatically sum the scores and apply a curve instead of having to do it all by hand.
I've done that too. The system was at a uni for a 3rd year module. The point with that is that everyone did the first two years, so the curve for those years was easy. For the 3rd year there were not enough people to grade on a curve, acurately, so the curve was derived from that cohort's second year results.
I did a spreadsheet, realised I had bugs, then rewrote it in AWK in about 1/4 of the time. But this isn't me being a smug git, so please read on...
To me, personally I don't like spreadsheets. This is not to say that they are bad, but I just do not get on well with them. Clearly many peopl do and so it's reasonable ot help them.
I have a vague sneaking suspicion that people on the whole will put in thousands of hours into using a bad tool to avoid learning a better one. For me, spreadsheets fall into the "bad tool" category. But them I'm rather fond of static strictness. I reailse the irony of that with my comment about AWK but bear with me.
It's like the old static versus dynamic debate. C++ provides rich static typing which is why I prefer it for large programs over a dynamically typed one like AWK. For small programs, it doesn't matter because I can track the types in my head. Likewise, the relational model provides static strictness over a model of data, which is why I prefer it for anything sizable which isn't a trivial list (most of my data is actually trivial lists).
And finally, spreadsheets are a nexus of dynamism: there's not even static checking of the program structure. It's this completely synamic 2D grid. This is not to say it's unusable: it's considerably better than the non grid equivalent of computed gotos because the 2D visual layout gets to make use of one's incredibly powerful visual cortex.
But ultimately, I find spreadsheets to be too dynamic for my taste.
OK, so at this point my thesis is that spreadsheets have dynamic types (they're not stringly typed---like AWK it looks a bit stringly typed but isn't, unlike TCL which explicitly is and takes that ti it's (il)logical conclusion---, I think they are actually properly dynamically typed on the whole even if the model is less powerful than Python)...
That got long. Spreadsheets have dynamic typing and dynamic structure.
Now, one can wade into innumerable debates into static versus dynamic typing on the internet. Mostly these are shouting matches. Even the accurate technical points (which is that static languages are a strict superset of dynamic ones) are more or less meaningless when it comes to practicalities.
Given the cognitive overhead, I don't thing spreadsheets scale well to even moderately sized programs (though a large spreadsheet is often large because it has a lot of data). Personally they don't even scale to small programs for me, but then not everyone thinks like I do and one's preference for staticness versus dynamism is really a function of how one thinks.
I've also ignored macros which allow one to indert statically typed, structured code into a spreasheet.
Ultimately though it doesn't matter what I think. I think people hammering on huge spreasheets are proven capable of reasoning in the type of way needed to write programs and I, personally think they would benefit from learning more structured languages. They won't though for a variety of reasons someof which are in fact perfectly good. So they will keep using vast spreasdsheets and AMD's hacking could prove very useful to them.