The famous interview questions of the last 30 years sort cleanly into three buckets in 2026: alive and asked daily, dead and only referenced as folklore, and still asked but on their way out. Knowing which is which is the most useful piece of preparation for a candidate who has read all the books and watched all the YouTube prep videos and is now trying to figure out what to actually drill before next week’s onsite.
This piece is the calibration map for the entire famous interview questions section. Every other piece in the section dives into a single famous question’s story; this one steps back and answers the practical question of where each question sits in 2026.
Live canon: still asked daily
Coding canon (still alive)
- FizzBuzz — alive at tier-2 tech and as a phone-screen warmup at some FAANG. Variants are more common than the literal version.
- Two Sum — universally asked at phone screens.
- Trapping rain water — alive at FAANG senior+ onsites. The optimization journey from O(n²) to O(n) is the actual signal.
- Median of two sorted arrays — alive at senior+ FAANG and AI lab interviews.
- Longest substring without repeating characters — universal at FAANG phone screens.
- Largest rectangle in histogram — alive at senior+ FAANG.
- Container with most water — alive at FAANG phone screens.
- Find median in a stream (two heaps) — alive at senior+ FAANG and at fintech.
- Word Search II (trie) — alive at FAANG senior+.
- LRU Cache design — universal at senior FAANG and tier-2 tech.
- Design URL shortener / Twitter / rate limiter — universal at senior+ system design rounds.
Behavioral canon (still alive)
- “Tell me about yourself” — universal across all roles, levels, industries.
- “Tell me about a time you failed” — universal at any level above junior.
- “What is your greatest weakness?” — still asked everywhere, despite being the most-mocked classic.
- “Why are you leaving your current job?” — universal when the candidate is currently employed.
- Conflict / disagreement stories — universal at senior+ levels.
- Amazon’s 16 Leadership Principles — universal in any Amazon loop. Several other companies have copied the format.
- Google’s “Googleyness” round — alive at Google.
Brainteaser canon (alive in finance, dead in tech)
- Ant on a clockface — alive at quant interviews.
- Pirates dividing gold — alive at quant interviews and senior tech reasoning rounds.
- Two children paradox / Tuesday boy — alive at quant interviews.
- Fair coin from biased — alive at quant interviews.
- Coin flip streaks (HT vs TT) — alive at quant interviews.
- 100 prisoners and the light bulb — alive at quant interviews.
- Secretary problem (1/e rule) — alive at quant interviews.
- Two envelopes paradox — alive at quant interviews.
- Pen-selling question (Wolf of Wall Street) — alive at investment banks and brokerage.
- Hostile / pressure-test interview format — alive at Wall Street; rare in tech.
Meta-questions (still alive)
- “Do you have questions for us?” — universal at the end of every loop.
- “Where do you see yourself in 5 years?” — declining but not dead. Still asked at large legacy enterprises and senior leadership tracks.
Dead canon: extinct or folklore-only
The retired Microsoft / Google brainteaser era
- “Why are manhole covers round?” — dead in tech. Still appears in nostalgia retrospectives but essentially never asked seriously.
- “How would you move Mount Fuji?” — dead. The question that named the era; the era itself is over.
- “How many golf balls fit in a school bus?” — dead since Google’s 2013 retreat. Still appears in PM-interview Fermi-estimation lists.
- “How many piano tuners are there in Chicago?” — dead in tech. Appears in PM and consulting interviews as Fermi estimation.
- The blender / shrunk-to-a-pencil — dead. Most absurd of the Microsoft era; nobody asks it seriously anymore.
- “How would you weigh a 747?” — dead in tech; still appears as estimation in some PM loops.
Coding questions in soft retreat
- Invert a binary tree — rarely asked as a primary question due to the Max Howell association. Sometimes appears as a low-pressure warmup.
- atoi / itoa — receding. Still useful as a spec-reading test, but newer parser-style questions have largely replaced it.
- Reverse a linked list — receding from primary onsite to phone-screen warmup.
The flipping zone: questions on the move
Some questions are in active transition — neither universally asked nor fully retired. These are the questions where preparation is most uncertain:
- Pure LeetCode-style algorithmic coding for senior frontend specialists. The Lyft frontend debate of 2024 reopened this argument. Some companies have differentiated; most have not. Frontend candidates should still prep algorithms but should also ask recruiters about format.
- Take-home assignments. Were rising fast in the late 2010s and early 2020s. AI coding tools have made take-homes harder to use as filters; some companies are abandoning them while others are tightening the rules. Direction in 2026 is unclear.
- Live debugging rounds. Increasingly common at companies that have lost faith in pure algorithmic interviews. Likely to keep growing.
- “What’s your greatest weakness?” — has been mocked enough that some companies have replaced it with “what would your last manager say you need to work on?”. Same question, less rehearsed phrasing.
- System design at the entry level. Used to be a senior+ format only. Increasingly appearing for mid-level roles, especially at infrastructure-heavy companies.
- AI-permitted coding rounds. Anthropic, some teams at Google, and a few AI labs allow candidates to use AI tools during coding interviews. Likely to spread; format details are still in flux.
Industry-by-industry breakdown
FAANG and tier-1 tech (Google, Meta, Apple, Amazon, Microsoft, Netflix)
LeetCode-style algorithmic coding (3 rounds), system design (1–2 rounds at senior+), structured behavioral (1–2 rounds). Brainteasers extinct. Take-homes rare. Hostile format rare.
AI labs (OpenAI, Anthropic, DeepMind, Cohere, Mistral)
Similar to FAANG plus ML-domain-specific rounds. Anthropic specifically has experimented with AI-permitted interviews. System design depth higher than typical FAANG.
Tier-2 tech (Databricks, Snowflake, Stripe, Atlassian, etc.)
Similar to FAANG with more variation. Stripe historically used take-homes; some others use live debugging. Algorithmic round is still standard.
Startups (early stage)
Highly variable. Some founders have read every interview-prep book and ape FAANG; others have rejected the format and use trial days, take-homes, or pure conversation. Ask the recruiter what to expect.
Wall Street (Goldman Sachs, Morgan Stanley, JPMorgan, Citi, BlackRock)
Probability brainteasers, mental math, behavioral, and the hostile / pressure-test format. Algorithmic coding less central than at FAANG. The Wolf of Wall Street pen-selling question still appears in sales-track loops.
Quant prop trading (Jane Street, Citadel, HRT, Two Sigma, Optiver)
Probability brainteasers, mental math, and at the more software-heavy firms (Jane Street, Citadel) coding rounds that resemble FAANG. Game theory and combinatorics questions are core. Hostile format common.
Hedge funds (Bridgewater, Renaissance, Point72, Millennium)
Mix of probability puzzles, behavioral, culture-fit. Bridgewater has its own famously elaborate process. Renaissance is essentially impossible to interview at without internal connections.
Government, defense, traditional enterprise
Conventional behavioral interviews, rarely cutting-edge. “Where do you see yourself in 5 years” still asked seriously. Algorithmic rigor lower than FAANG.
What to drill in 2026
For a tech-track candidate without specifics about the company:
- LeetCode mediums via Blind 75 or Neetcode 150.
- System design via Alex Xu’s books and Designing Data-Intensive Applications.
- “Tell me about yourself” + “tell me about a time you failed” + “why are you leaving” — drilled to the structure level, not memorized word-for-word.
- Amazon LPs if you are interviewing at Amazon; otherwise standard behavioral.
For a quant-track candidate:
- Mark Joshi’s Quant Job Interview Questions and Answers for probability puzzles.
- Mental math drills with a daily timer.
- Pen-and-paper combinatorics and game-theory practice.
- Hostile-interview practice with a friend.
For both: the live canon will still be asked. The dead canon will not. The flipping zone is where to ask recruiters and adapt as you learn what the specific company uses.
Frequently Asked Questions
Are brainteasers really dead in tech?
The pure brainteasers (manhole, Mount Fuji, golf balls) are essentially extinct. Estimation questions still appear in PM interviews. Probability puzzles are alive only at quant firms.
Do I still need to grind LeetCode for FAANG?
Yes. LeetCode-style coding is still the dominant filter at FAANG and shows no sign of disappearing in 2026. Blind 75 plus targeted topic depth is the bar.
What’s the most durable interview format?
The behavioral round. The questions and frameworks have been stable for 25 years. If you only have time to prepare one format perfectly, this is the highest-leverage choice.
How do I find out which questions a specific company asks?
Three sources: the recruiter (will tell you the format if asked), Glassdoor and Levels.fyi (leaked questions are often accurate), and the company’s own engineering blog (sometimes describes the interview process).
Will the format change again before 2030?
Probably yes. The trajectory of tech interview formats has been to evolve every 5–10 years. AI-assisted interviews and a possible retreat from LeetCode are the most likely next shifts.