Not just in one HN'ers opinion.
The thing is: Excel is one part data store and one part programming interface. The programming interface just happens to be a part of the data store. This leads people who have been told they're not allowed to program to solve the problem with the best tool they can hold: Excel.
I once got an email in my inbox that had the subject "Help: recommendations for machine to run Excel". The body described a person within a business unit who had a laptop, ca. 2018, with a then current I7, 32G of RAM, dual raid10 SSDs, and it had one job: Open, edit, and evaluate one 5+GB Excel workbook.
This workbook took twenty minutes to open.
It's so bad Microsoft actually gave some level of proper functional programming to Excel: https://www.theregister.com/2020/12/04/microsoft_excel_lambd...
What happens is this pattern where more applicable programming models are taken from business folk, who need a way to do some level of broad, tabular accounting usually. These individuals don't have Python, don't know Python, or web applications, etc. What they have is Excel.
Excel lets you do something very easily that no other thing in existence lets you do: Real time (ish) calculations against very large sets of data. I can write one cell, `=AVERAGE($A9:$A99)` and get the running average for 90 rows in a table. Imagine the work that would go into doing that in a language like Python or Go, plus the scaffolding to make it work in a webapp.
And so when a business unit says "we need a way to do X", inevitably someone says "there's no budget for that" and someone eventually finds their way to Excel. Then they build the thing in Excel and when someone realizes they have a monster on their hands someone inevitably says "this should be a webapp" at which point nobody has the funding to maintain that webapp plus its infrastructure, plus handling the implicit and explicit logic that comes along with the mixture of data, code, and everything else in an Excel spreadsheet.