The one-liner Python writes when it's showing off — step 1 of 9
The one-liner AI reaches for nine times out of ten
Watch what happens the next time you ask Cursor for a list of email
addresses from these users or the prices in dollars instead of
cents or every URL in lowercase. It will not write a four-line
for loop with .append. It will write one line. That line is a
list comprehension.
This is the most common single-line construct in modern Python, and roughly half of "I can't read AI's code" complaints from non-engineers trace back to this exact syntax. Once you've internalized the shape, the rest of the language opens up.
The mental model: read it like English
A list comprehension is built from three pieces, in order:
[ expression for item in iterable ]
^new-value ^source-loop
Read it left-to-right, out loud if it helps:
"a new list, where each item is
expression, for everyiteminiterable."
That's it. No new logic. It's the same for loop you already know,
collapsed onto one line and wrapped in [ ] to mark "this whole
thing is a list."
The long way and the short way, side by side
The editor on the right has both forms doing exactly the same job:
# the long way
with_tax = []
for p in prices:
with_tax.append(p * 1.1)
# the comprehension
with_tax_short = [p * 1.1 for p in prices]
The long way is what you'd write if you'd never seen Python before.
Three lines of setup, loop, append. The comprehension says the same
thing in twelve characters less, with one less moving part — no
explicit empty list to start, no .append call.
When you read the comprehension, the trick is to read the right side
first: for p in prices tells you the source. Then read the left:
p * 1.1 tells you what to do with each p. The brackets tell you
the result is a list.
Why AI almost always writes it this way
Three reasons Cursor reaches for comprehensions:
- Less code to maintain. One line vs. four.
- No empty-list-and-append boilerplate. Less to get wrong.
- Faster at runtime. Comprehensions are slightly faster than the manual loop because Python optimizes them internally. For tiny lists this doesn't matter; for big ones it does.
You'll start writing them yourself within a week of seeing them. They become muscle memory.
Where AI gets comprehensions wrong
The single most common mistake: nesting them so deep nobody can read them. Cursor sometimes writes:
emails = [u["email"] for team in org for u in team["members"] if u["active"]]
That's a valid comprehension, but it crams a lot into one line. When
you see one with two for clauses or an if, break it into a real
loop when you're reading. Six readable lines beat one clever line
every time.
Run the editor. Both lines print the same list — [11.0, 22.0, 33.0].
Same output, two flavors of syntax.
The one-liner Python writes when it's showing off — step 1 of 9
The one-liner AI reaches for nine times out of ten
Watch what happens the next time you ask Cursor for a list of email
addresses from these users or the prices in dollars instead of
cents or every URL in lowercase. It will not write a four-line
for loop with .append. It will write one line. That line is a
list comprehension.
This is the most common single-line construct in modern Python, and roughly half of "I can't read AI's code" complaints from non-engineers trace back to this exact syntax. Once you've internalized the shape, the rest of the language opens up.
The mental model: read it like English
A list comprehension is built from three pieces, in order:
[ expression for item in iterable ]
^new-value ^source-loop
Read it left-to-right, out loud if it helps:
"a new list, where each item is
expression, for everyiteminiterable."
That's it. No new logic. It's the same for loop you already know,
collapsed onto one line and wrapped in [ ] to mark "this whole
thing is a list."
The long way and the short way, side by side
The editor on the right has both forms doing exactly the same job:
# the long way
with_tax = []
for p in prices:
with_tax.append(p * 1.1)
# the comprehension
with_tax_short = [p * 1.1 for p in prices]
The long way is what you'd write if you'd never seen Python before.
Three lines of setup, loop, append. The comprehension says the same
thing in twelve characters less, with one less moving part — no
explicit empty list to start, no .append call.
When you read the comprehension, the trick is to read the right side
first: for p in prices tells you the source. Then read the left:
p * 1.1 tells you what to do with each p. The brackets tell you
the result is a list.
Why AI almost always writes it this way
Three reasons Cursor reaches for comprehensions:
- Less code to maintain. One line vs. four.
- No empty-list-and-append boilerplate. Less to get wrong.
- Faster at runtime. Comprehensions are slightly faster than the manual loop because Python optimizes them internally. For tiny lists this doesn't matter; for big ones it does.
You'll start writing them yourself within a week of seeing them. They become muscle memory.
Where AI gets comprehensions wrong
The single most common mistake: nesting them so deep nobody can read them. Cursor sometimes writes:
emails = [u["email"] for team in org for u in team["members"] if u["active"]]
That's a valid comprehension, but it crams a lot into one line. When
you see one with two for clauses or an if, break it into a real
loop when you're reading. Six readable lines beat one clever line
every time.
Run the editor. Both lines print the same list — [11.0, 22.0, 33.0].
Same output, two flavors of syntax.