Type issues with Callables

• Are you reporting a bug, or opening a feature request?
  bug

• Please insert below the code you are checking with mypy

from typing import Optional, Iterable, TypeVar, Callable

T = TypeVar('T')


def is_even(n: int) -> bool:
    return n % 2 == 0


# equivalent to https://ruby-doc.org/core-2.5.0/Enumerable.html#method-i-detect
def detect(function: Callable[[T], bool],
           iterable: Iterable[T]) -> Optional[T]:
    for element in iterable:
        if function(element):
            return element
    return None


numbers = [1, 2, 3, 4, 5]

# using a lambda
print(detect(lambda n: n % 2 == 0, numbers))

# using a function
print(detect(is_even, numbers))

• What is the actual behavior/output?

$ mypy detect.py 
detect.py:22: error: Unsupported operand types for % ("object" and "int")
detect.py:25: error: Argument 1 to "detect" has incompatible type "Callable[[int], bool]"; expected "Callable[[object], bool]"
Found 2 errors in 1 file (checked 1 source file)

• What is the behavior/output you expect?
  I was expecting the types to be inferred and the above code to typecheck.

• What are the versions of mypy and Python you are using?

$ mypy --version
mypy 0.770
$ python --version
Python 3.7.7

• Do you see the same issue after installing mypy from Git master?
  yes

$ mypy detect.py
detect.py:22: error: Unsupported operand types for % ("object" and "int")
detect.py:25: error: Argument 1 to "detect" has incompatible type "Callable[[int], bool]"; expected "Callable[[object], bool]"
Found 2 errors in 1 file (checked 1 source file)
$ mypy --version
mypy 0.770+dev.22c67daac7d2b28f33d4e9c7bee4f1a06e61a3e5

• What are the mypy flags you are using? (For example --strict-optional)
  None.

• If mypy crashed with a traceback, please paste the full traceback below.
  It didn't crash

Start by running the supplied detect.py reproducer with the reported mypy versions and inspect how Callable and TypeVar inference handle the lambda and is_even arguments. Trace the relevant type-checking behavior and add regression coverage where appropriate; done means the example typechecks without the two reported errors.