Nobody is hiring vibe coders. Sorry
What this video is
⚡ a 12-minute video, readable in 60 secondsA Traversy Media host argues that coding interviews are not disappearing in the AI era, they are multiplying and getting harder, while still testing the same underlying judgment they always did. His case blends different kinds of grounding: a personal anecdote about recently interviewing with 11 companies, cited industry data from CoderPad's hiring reports and Anthropic's published candidate AI policy, and claims about OpenAI's and Meta's interview practices that are not tied to a named source. He also asserts, without citing a source, that technical interviews have nearly doubled in the US since 2023. The video closes urging viewers to keep their fundamentals sharp and get fluent with AI tools, alongside a promotion for the hosts' own platform, Start.dev.
Thesis [12:01]: coding interviews aren't dead, there are more of them than ever, and they still test what they always tested, whether you know what you're doing, which AI didn't change.
Key takeaways
- Argument [02:56]: Meta built an AI-assisted interview round that, by most accounts, is harder than its classic round even though it encourages AI use.
- Argument [03:05]: the speaker claims the companies closest to the AI models test candidates the hardest.
- Argument [05:14]: many interviewers let candidates choose their own programming language because syntax was never the part that mattered.
- Argument [05:22]: interviewers grade judgment, whether a candidate can think through a problem, justify decisions, weigh trade-offs, and identify where a solution breaks at scale.
+ 42 more takeaways
- Argument [06:44]: watching whether a candidate catches obvious issues or glazes over them is described as a great way to evaluate them.
- Argument [07:10]: developers have complained about interviews being too abstract and too much memorization since long before AI, and for many years nothing changed.
- Argument [07:31]: companies are starting to swap single function puzzles for real world simulations on multi-file codebases with realistic debugging instead of greenfield toy projects.
- Argument [08:01]: the presenter calls this shift the first change in years he would call progress in the interview process.
- Argument [08:40]: the presenter frames 70 years of interview evolution as a stack, mechanical coding at the bottom, implementation in the middle, judgment and trade-offs at the top, where each era of tooling deletes the bottom layer, and AI is now deleting the layer of producing actual code.
- Argument [10:41]: the presenter states he would not hire a developer who refuses to leverage AI tooling, outside of some specific use cases.
- Argument [11:00]: the speaker says it's getting irresponsible not to use AI, and says resistance to AI is itself now becoming a problem.
- Argument [11:45]: the speaker says we're past the 'AI is just a tool' phase and people need to adapt.
- Sourced claim (CoderPad) [01:36]: technical assessments are up 48% globally since mid-2023.
- Sourced claim (CoderPad) [01:36]: 43% of hiring teams still run classic algorithmic assessments.
- Sourced claim (CoderPad) [01:52]: the remaining 57% of companies shifted the technical bar to system design interviews, pair programming, and real-world debugging instead of dropping it.
- Sourced claim (Anthropic) [02:40]: Anthropic's official candidate AI policy states there is no AI assistance during live interviews unless they indicate otherwise.
- Sourced claim (Anthropic) [02:43]: the guidance's example prep prompt reads, 'Please review my resume and the job description. Identify the experiences I should highlight in my application responses that align most with the job requirements.'
- Sourced claim (Anthropic) [02:43]: for take-home assessments, candidates should complete them without Claude unless Anthropic indicates otherwise.
- Sourced claim (Anthropic) [02:43]: for interview prep, candidates may use Claude to research the company, practice answers, and prepare questions.
- Sourced claim (Anthropic) [02:43]: during live interviews no AI assistance is allowed unless indicated otherwise, and candidates needing accommodations should tell their recruiter early.
- Sourced claim (Anthropic) [02:49]: Anthropic states it wants to see how candidates think and reason through problems in real time.
- Sourced claim (CoderPad, State of Tech Hiring 2026) [08:13]: roughly 34% of companies ban AI entirely in interviews.
- Sourced claim (CoderPad, State of Tech Hiring 2026) [08:13]: 46% allow AI broadly or with constraints.
- Sourced claim (CoderPad, State of Tech Hiring 2026) [08:13]: 20% decide case by case.
- Unsourced claim [00:10]: technical interviews have nearly doubled in the US since 2023 and are becoming more challenging, no source is named for this figure.
- Unsourced claim [02:21]: OpenAI's most commonly reported interview question is implementing an LRU cache from scratch live while explaining your reasoning, no source is named for how this was determined.
- Unsourced claim [06:19]: candidates in Meta's AI-assisted round reportedly say the AI models seem dumber during the interview than during practice the day before.
- Unsourced claim [06:31]: the same models and prompts reportedly write buggy code and fail to point out obvious mistakes during the interview.
- Unsourced claim [06:38]: the speaker says it is not confirmed by Meta that they purposely make their models dumber, but says it wouldn't surprise him.
- Unsourced claim [08:21]: where AI is allowed, some companies reportedly ask candidates to hand over their prompt history and deep-grill them on take-home projects.
- Unsourced claim [10:11]: 74% of developers say landing a tech job is difficult while 78% of tech leaders say they can't find qualified candidates, no source is named for this figure.
- Anecdote [00:49]: the host says he recently interviewed with 11 different companies, and every one he got far enough with led to a technical interview.
- Anecdote [01:01]: those interviews were a mix of coding challenges, system design questions, and questions tied to each company's specific niche.
- Anecdote [01:14]: two of the 11 interviews used the classic 'Valid Sudoku' coding problem.
- Anecdote [05:47]: the speaker estimates fewer than 50% of the interviews he took were actual coding interviews, with most being scratch-pad or whiteboard walkthroughs.
- Anecdote [05:56]: the best interviews the speaker conducted involved candidates having a real conversation and walking through their reasoning rather than just giving the answer.
- Anecdote [11:17]: in 2017 the speaker hired his first developer for a project that had scaled up and needed team members.
- Anecdote [11:26]: that developer refused to move out of the PHP stack and complained about the frameworks being used.
- Anecdote [11:33]: the developer's claim was that real developers should write their own web server from scratch, and that frameworks like Django and React were just for noobs.
- On-screen example [07:52]: a demoed bug in applyDiscount.ts lets stacked coupon percentages exceed a 50% cap, causing a test to fail with 'expected 50 received 65.'
- On-screen example [07:53]: the shown fix changes the line to pct = Math.min(pct + rule.percent, MAX_PCT), which makes all 8 tests pass.
- Prediction (no timeframe stated) [11:49]: the speaker says people will be graded not only on ability but also on AI fluency.
- Ask [03:37]: the hosts promote Start.dev, their interactive learning platform with bite-sized lessons and an in-browser code editor.
- Ask [04:26]: Start.dev is free to start and offers 20% off premium access via a coupon code in the video description.
- Ask [04:33]: the coupon code 'TRAVERSY' gives 20% off Start.dev premium.
- Counterpoints: not stated.
Shown on screen — grab and go
PROMPTAnthropic's example resume-highlighting AI prompt
Please review my resume and the job description. Identify the experiences I should highlight in my application responses that align most with the job requirements.
Quoted from the Anthropic candidate-AI-guidance page shown on screen at 02:43; only spacing was restored from crushed OCR. shown at 2:43
PROMPTAnthropic's example interview study-guide prompt
Create a study guide for interviewing for this job. Outline key topics I should review, including AI safety concepts, Anthropic's research focus, and typical technical or behavioral questions I might encounter.
Quoted from the same Anthropic guidance page, confirmed by a second clean OCR read at 02:49. shown at 2:49
CODEThe applyDiscount.ts coupon-stacking cap fix
export function applyDiscount(cart: Cart, codes: Code[]) {
const rules = loadRules(cart.region);
let pct = 0;
for (const code of codes) {
const rule = rules.get(code);
if (!rule) continue;
// NOTE: legacy stacking path - see pricing/rules.ts
pct = Math.min(pct + rule.percent, MAX_PCT);
[...]
return subtotal(cart) * (1 - pct);Reconstructed from the applyDiscount.ts editor shown 07:32-07:56 using the final passing version at 07:56; lines obscured behind the terminal panel are marked [...]. shown at 7:56
CODEWhiteboard pseudocode for reversing a linked list
prev = null while (curr) next = curr.next curr.next = prev prev = curr, curr = next → O(n) time, O(1) space
Transcribed from the whiteboard drawing shown 07:12-07:22; line breaks inferred from the diagram layout, no words altered beyond spacing. shown at 7:16
QUOTETweet quote comparing coding to horse grooming
Learning to code today is like learning horse grooming in 1910. — Sunny Marchetti, @futurestackr
Quoted from the tweet screenshot shown at 00:06, with attribution to @futurestackr visible on screen. shown at 0:06
How this brief was shaped: Trend Commentary / Opinion Essay · confidence Low
The opening frames a viral claim, coding interviews are dead, then pivots to data, technical interviews nearly doubled since 2023, and backs it with a personal case study of interviewing at 11 companies, which is the thesis plus claims-sorted-by-evidence structure of an opinion essay. OCR of Anthropic's public candidate AI guidance page supports the argument about how interviews are actually run now rather than showing code or a tutorial.
The lens sets this brief's structure, never its facts — every claim is held to the same citation and fact-check standard.
Their links, sorted & clickable
🔗 Other links1
Join start.dev todayWas this brief useful?