All Epigrams are One-Sided and Some are Overused.

An eminent speaker's explaining their theory.
You listen. You quickly get angry and weary.
You study the model they tried to construct
You see it's nonsensical — wrong — upgefucked.
The components are nebulous (they admit that).
The numerical values were pulled from a hat.
Your colleague's confused and they clearly conflated
Two concepts not even remotely related.

The assumptions built into their model stochastic
Are false, inconsistent, far-fetched, and fantastic.
The argument's piffle, the logic's unsound,
The causal direction's the other way round.
But if you should dare to point some of this out
Some boob in the crowd will undoubtedly shout
That most overused instance of threadbare excuses:
"All models are wrong, but some have their uses."

Among all the quotations that engineers love,
I don't think there are any I'm more tired of.
The original point was important and real,
And the quote has immense aphoristic appeal.
But too often it's used as a blanket to coddle
A scientist shilling a meaningless model.
George Box's quotation is clever. It's not
A justification for modelling rot.

And certainly there is no reason to stop
Denouncing that foul intellectual slop
'Cause some idiot quotes Box's smart-alec rule,
The only advice they remember from school,
Some pompous, illiterate, arrogant fool,
Who wishes to seem both hard-boiled and cool,
Untaught, obscurantist, Bullwinkle-the-Moose-ful
"All models are wrong but some models are useful."

This is part of the collection Verses for the Information Age by Ernest Davis