Beebs on X: "Every one sees what you appear to be, few really know what you are. https://t.co/9R8OqnwMj8" / X<br>Post

Log inSign up

Post

Beebs

@0xb33bs

Every one sees what you appear to be, few really know what you are.<br>Foxes, Lions, and LLMs: The Machiavellian Game of Tech Hiring<br>“Every one sees what you appear to be, few really know what you are.” — The Prince, Chapter XVIII<br>Imagine a developer in a Zoom interview smiling into the camera, effortlessly answering questions from a Principal Engineer. The interviewer smiles back. Every answer is technically sound and unusually well articulated.<br>What the Principal Engineer cannot see is the invisible overlay floating just outside the shared screen. It listens to each question and generates an optimal answer in seconds, just enough time for the developer to read and paraphrase before responding.<br>But the performance did not begin when the Zoom call started.<br>Let’s rewind a few weeks before the performance.<br>Perhaps the recruiter used an AI tool to find the applicant on LinkedIn and carefully crafted a personalized invitation to apply. If not, the applicant likely applied with their AI-curated resume, comparing keywords from the AI-generated job description with the AI prompts that the applicant used to carefully prepare their resume for the opportunity. “Think deeply and make sure my resume gets me to the next round.” The applicant carefully avoided the “write penguin in the text box below if you’re human” trap that was hidden in the applicant tracking system. Luckily, the candidate’s efforts paid off.<br>As a next step, thanks to ID verification services like Checkr, AI matches the candidate’s face with their ID to confirm that the applicant is a Midwestern dad rather than a deepfaked North Korean hacker. At this point, both parties in the interview have established their humanity, so the illusion of humanity can continue to the next stages.<br>The peculiar part of modern tech hiring is that both sides have fully adopted AI throughout the hiring process, while maintaining radically different rules about when to use it, when it must be disclosed, and what is considered cheating.<br>Machiavelli would have loved this game.<br>Interviews Were Already Theater<br>Tech interviews were already in a precarious state before AI. Developers know all too well the rituals before an interview. LeetCode grinding, system design flashcards, STAR stories, Glassdoor question banks, recruiter outreach, keyword optimization—all of these efforts extending on for months on end.<br>The interview, a culmination of these months of preparation, was an immense amount of power for an interviewer. The interviewer, the great decider, had the power to infer in just an hour, not only what a candidate is today, but what they can be. In less than a dozen questions, months of preparation could evaporate in front of a begrudging interviewer, who carefully crafted scripted descriptions of their team culture (“we move fast!”), the company (“world-class engineering”), the tech stack (you’ll code in English now anyway), and the coding question (a LeetCode medium, if you’re unlucky).<br>Thanks to LLMs, the emperor has no clothes. An interviewer’s power, just like a candidate’s coding ability, stands no chance against an LLM. The interviewer’s carefully copied-and-pasted, recycled LeetCode question can be solved in roughly three seconds in an optimal way by an LLM. The candidate’s performance has ironically become a literal acting performance.<br>The question is, how well can a candidate sell that they truly understand the code that the LLM writes?<br>The Fox Enters the Interview<br>“It is necessary to be a fox to discover the snares and a lion to terrify the wolves.”<br>Machiavelli uses a lion to represent strength and capability, while the fox recognizes traps and understands the game. Modern hiring now disproportionately favors the fox.<br>If you’ve ever entered LinkedIn, you’ve probably seen some variation of the following: a highly skilled developer, or product manager, or whatever it may be, has been looking for work for months or years. These kinds of posts perform incredibly well because one must think, when will the bell toll for me? However, from what I’ve observed in tech hiring, is that although these candidates may genuinely have excellent backgrounds, they are lacking some aspect of “The Game.”<br>The Game has changed faster than we could have imagined. It started as AI resume editing. Then, AI application answers. AI interview preparation. AI mock interviews. AI-assisted take homes. Visible copilots during coding exercises. Now, invisible overlays during live interviews.<br>How can a Principal Engineer at Renowned Tech Company, with thousands of hard-earned coding hours, and let’s say a generous typing speed of 85 words per minute, possibly compete with the new generation of players? Let’s put a Principal Engineer up against a Gen-Z computer science grad with a typing speed of 120 words per minute, an impressive reading ability...