Atlas · Interviewing · 12 entries
Interviewing
I've written about interviewing for most of my career. It's truly one of the worst corners of our industry, and I ultimately lost faith in the whole process about ten years ago.
Most of my posts here are for interviewers: posts written to help you figure out how to evaluate candidates. Unsurprisingly, my most famous of these posts by far is for interviewees — Get That Job At Google is a guide on how to be your best in interviews.
I believe interviewing is about to change fundamentally. We dropped phone screens as soon as we no longer needed them, and we switched to coding on laptops (instead of whiteboards) as soon as we could. So interviewing has already changed shape in two major ways while I've been doing it. My 2026 post, The Last Technical Interview, predicts that bringing in AI will necessitate switching to higher-signal evaluation processes. Time will tell.
- ★ Essentials
- 👍 Good read
- 💩 Not worth it
- 🤓 Language nerd
- 😄 Humor
- 📚 Story / fiction
- 🕹️ Gaming
- 🔮 Called it
- 🤡 Whiffed it
-
2026
The Last Technical Interview
After 35 years of conducting technical interviews — Amazon Bar Raiser, Google Hiring Committee — Steve argues the interview loop is finally dying on its own, strained past breaking by AI-flooded résumés, cheating, and roles that now change faster than any question bank. The replacement he sketches is the campfire: paid, real work on a real codebase, with a portable record the candidate keeps whether or not they're hired.
-
2018
Get That Job at Grab
Half labor-market broadside (global demand for engineers has outrun supply, so go look around), half interview-prep advice (breadth-first beats depth-first, one mulligan per candidate), with a closing parable Steve calls the Steve Buscemi Test about candidates who literally couldn't count to a million.
-
2008
Done, and Gets Things Smart
An essay built on the Dunning-Kruger Effect, with Steve as the example: incompetent people overestimate themselves and fail to recognise skill in others — which makes "hire smart people" circular. Classic interviews, Steve argues, mostly clone the interviewer. What you want instead is rarer: the seed engineer who finishes absurdly fast and improves the systems and the culture while they are in there.
-
2008
Get That Job at Google
Steve finally gave in and wrote the interview-prep post he had been dodging for years, knowing it would draw flames. Underneath the comedy is genuine service journalism: study a data-structures book, run mock whiteboard interviews, know Big-O cold, know your graphs — and understand that a rejection from a high-false-negative-rate company is often just noise, not a verdict on you.
-
2007
Ten Tips for a (Slightly) Less Awful Resume
After years of reading bad programmer resumes from inside Google and Amazon, Steve dumps a working pass at why the resumes are so uniformly awful, and the ten things he wishes candidates would stop doing. Tone is deadpan, examples are real, the advice is unsentimental.
-
2006
The Truth About Interviewing
Steve takes the comment seriously and writes the post he probably should have written first — what technical interviewers are actually doing, what they cannot ever know, and why the process is the way it is anyway. The title is meant to be partial and biased on purpose.
-
2005
The Miracle Interview
Steve runs through the loop in real-time: a string-permutations fumble, a phone-numbers question solved by writing a small C program instead of using
grep, the candidate scrunching his diagrams into the rightmost two inches of a wide- open whiteboard. Then Steve stops the interview, narrates the bullet points of failure on the board itself, and explains what needs to change. The candidate listens, nods, and proceeds to smoke a parking-lot OO design, the entire operating-systems battery, and a polymorphic AST evaluator built up from nothing. Vote: inclined to hire. Everyone else on the loop voted no. Miracle cancelled. -
2005
Practicing Programming
Drawing on his years as a self-taught guitarist — and on the hard lesson that "practice makes perfect" is wrong, because perfect practice makes perfect — Steve argues that merely doing your job is not practice. Great engineers are great because they train deliberately and constantly. The essay closes with twelve concrete practice drills you can start today.
-
2005
Why Phone Screens Matter
Steve walks through the on-site that triggered the piece. SDE-2 candidate, fifteen years of C++ on Unix and Windows, two phone screeners voted to fly him in. He arrives, fails the regex question, doesn't know
man, can't traverse a graph or compute the depth of a binary tree, and thinks the inverse of 2N issqrt(n). Every interviewer on the loop votes against. The essay's argument is about the screening pipeline, which silently treated "fifteen years of C++" as proof of basics it never established. -
2005
Age of the Racecar Driver
The piece opens on a phone screen with a candidate Steve calls Mr. Earnhardt: a CS-adjacent degree, a long J2EE résumé, and no idea how a hashtable works. Steve uses the screen as a way into the longer argument. Software is becoming, like medicine, a field of specialised roles that look superficially similar but require very different training — doctors, nurses, pharmacists, orderlies, dental hygienists. Software has hospitals full of "doctors" all the way down. The racecar-driver / racecar-engineer split is Steve's pet metaphor for the obvious next step. Both are valuable; both are hard; neither has to be both at once.
-
2004
Google's Secret Weapon
The piece begins as a write-up of "How to Win Big at Recruiting" and then abandons that title halfway through, because the candidates Steve has just been talking with have made it clear that a few tips aren't going to be enough. Google's recruiting is no longer a difference in magnitude. Smart people now make the pilgrimage to Google; Google spends most of its recruiting time turning candidates away. Steve writes down in 2005 what most recruiting orgs would spend the next decade learning the hard way.
-
2004
The Five Essential Phone-Screen Questions
After studying roughly a hundred failed phone screens, Steve found two patterns: the screener let the candidate do all the talking, and one-trick ponies only know one trick. His fix is five universal litmus-test areas — coding, OO design, scripting and regexes, data structures, and bits and bytes — quick to ask, brutal to fail, and a reliable way to spot a doomed interview before it wastes a full loop.
