For a stretch of about ten years from the mid-1990s through the mid-2000s, the manhole cover question was the most famous interview question in tech. Microsoft was its home — every Microsoft hiring retrospective from that era mentions it. The question is short and innocent-looking:
“Why are manhole covers round?”
Five words. No code. No data structures. Just an observation about the world and an invitation to explain it. By the time William Poundstone published How Would You Move Mount Fuji? in 2003, the manhole cover question had already passed into industry folklore. By the time a generation of CS students had read Poundstone’s book, the question was so universally rehearsed that asking it told you nothing about the candidate. That was the entire arc of the brainteaser era in tech, compressed into one question.
The canonical answers
There are several “right” answers, and a competent interview response is one that produces multiple of them and reasons about which is most fundamental.
- A round cover cannot fall into its own hole. A square cover, rotated 45 degrees and tipped on its diagonal, fits through the square opening because the diagonal of a square is longer than its side. A circle has constant width in every direction, so no rotation of a round cover lets it slip through a round hole. This is usually considered the answer because it cites a real engineering hazard — a falling manhole cover into the hole below could kill a worker or destroy equipment.
- A round cover does not need to be aligned to fit. A square cover has to be oriented before being lowered onto the hole. A round cover drops in regardless of rotation, which is a significant labor saving when you have crews installing covers all day.
- A round cover is easier to move. A heavy round cover can be rolled on its edge to and from a workbench. A square cover has to be lifted and carried. For a 250-pound iron cover, this matters.
- Less material for the same minimum cross-section. A circle is the shape with the smallest perimeter for a given enclosed area. If the constraint is “the hole has to be at least X inches across so a worker can fit through”, a round shape uses less iron and is cheaper.
- Stress distribution. A round cover distributes load uniformly around its perimeter. A square cover has corners that concentrate stress and tend to crack first.
The “good answer” in interview conventions was the first one (cannot fall in), followed by structured enumeration of the others. The “great answer” was to first ask “are they always round?” — manhole covers are not always round; many European designs use squares, rectangles, or hexagons — and to then walk through the trade-offs of each shape rather than presenting roundness as obviously correct. That clarification-first habit was what the interview was trying to elicit.
Why Microsoft in particular
The question was not invented by Microsoft. It had circulated in puzzle circles for decades before the 1990s; the manhole cover and the constant-width Reuleaux triangle had been part of recreational mathematics since the 19th century. What Microsoft did was elevate it to the canonical opening question of a particular kind of interview — the “let’s see how you think” round, where the actual problem was less important than the candidate’s process for attacking it.
Microsoft’s hiring philosophy in that era was heavily influenced by Bill Gates personally and by hiring memos that emphasized raw cognitive ability. The premise was that a smart candidate could be taught any specific technology, but a candidate who could not reason about an unfamiliar problem would always be stuck inside the patterns they already knew. The brainteaser was the test for “can this person attack a problem they have not seen before”. That was a coherent theory in 1995 when the question was novel.
The fall
The manhole cover question stopped working for the same reason every famous question stops working: it leaked. By 1998, every CS student had heard about it. By 2000, it had been written up in newspapers and industry magazines. By 2003, Poundstone’s book had collected it, and asking it was a tell that you had not updated your interview script in five years. By the late 2000s, asking the manhole question seriously was almost a self-deprecating joke — interviewers would lead with it precisely because everyone knew the answers, as a way to put the candidate at ease.
The deeper issue was empirical. When companies started measuring whether brainteaser performance predicted job performance, the answer was: not really. Google’s Laszlo Bock famously told the New York Times in 2013 that “brainteasers are a complete waste of time” — a public retreat from a hiring tradition Google had also embraced for years. Microsoft never made an equally explicit retraction, but the brainteaser style was quietly de-emphasized through the 2000s and 2010s, replaced by behavioral and coding rounds.
What the question was actually probing
A modern interviewer reading the brainteaser literature charitably can identify what the question was supposed to test. It was not, as the cynical reading would have it, “did you read this book”. It was three things layered together:
- Tolerance for ambiguity. The question gives you no constraints — no specifications about geography, time period, weight, material, or use case. A candidate who freezes when the problem is under-specified is going to struggle in any role where requirements are negotiated rather than handed down.
- Quality of clarifying questions. “Are they always round?” is a great clarifying question. So is “in what country / city?” The candidate’s first move signals whether they will ask before assuming.
- Structured enumeration. Producing five answers and ranking them by importance is a different skill from producing one answer and stopping. The interviewer was watching for whether the candidate kept generating after the first plausible answer.
None of this is unreasonable to want to test. The problem is just that the manhole cover question, after enough exposure, stopped testing any of these things — it tested whether the candidate had memorized a list of answers from a book.
Is anyone still asking the manhole cover question in 2026?
Almost no one in tech. Asking it in 2026 would be a strong signal that the interviewer has not updated their process in 20 years and has not been paying attention to industry research on what predicts engineering performance. There is one mild exception: some product manager loops still use Fermi-estimation questions (“how many manhole covers are there in Manhattan?”) as a way to test market-sizing intuition. That family of questions is alive and useful in PM interviews, but it is a different question with a different purpose.
Wall Street and quant firms never used the manhole cover question — their tradition uses probability puzzles instead, where the underlying skill (computing expected values quickly) is actually job-relevant. So the question is essentially extinct in serious technical hiring as of 2026, surviving only as a cultural reference.
What replaced it
The function the brainteaser used to serve — testing problem-solving on an unfamiliar task — is now mostly served by two replacement formats. The first is the LeetCode-medium coding problem, where the unfamiliarity comes from the specific algorithmic combination rather than from the problem domain. The second is the system design round, which is fundamentally a structured-ambiguity exercise: the interviewer gives a vague prompt (“design Twitter”), and the candidate’s job is to scope, clarify, and propose. System design retains the open-ended quality that made brainteasers attractive in the 1990s, while testing skills that are obviously relevant to building actual software.
Frequently Asked Questions
Is the manhole cover always round?
No. Many European cities use rectangular or hexagonal manhole covers. The “round so it cannot fall in” reasoning applies only to the specific case of a round opening with a round cover. Hexagonal covers do not fall through hexagonal openings either. The American convention favored round, which is why the puzzle assumes roundness in its setup.
Did Microsoft invent the manhole cover question?
No, the puzzle predates Microsoft by many decades. Microsoft popularized it as an interview question in the 1990s, but the underlying observation about constant-width shapes is part of recreational mathematics going back to the 19th century.
Should I prepare for brainteasers in 2026?
For pure tech interviews at FAANG, AI labs, or fintech, no. They are essentially extinct. For Wall Street quant interviews, yes — but the questions there are probability and game theory, not “why are manhole covers round”. For PM interviews, prepare for Fermi estimation, which is a related but distinct genre.
What is the best answer if I do get asked it?
Lead with “are they always round?” to signal that you check premises. Then give the not-falling-in answer. Then enumerate two or three other reasons (alignment, rolling, material). Stop before producing six trivia answers from memory. The signal is composure plus a clarifying-first habit, not whether you can recite the entire list.