December 27, 2025
Glasses-gate: ChatGPT caught in the glare
Reflections and rantings from a system design interviewer
Interviewer shames AI parrots; commenters call hiring a hot mess
TLDR: An interviewer scolds candidates for AI parroting and not researching companies, urging authentic answers and customer focus. Commenters clap back, saying hiring signals are inconsistent—some interviewers even demand more AI—while others argue revenue comes from marketing and UX, not fancy system design, exposing confusion in how tech hires are judged.
A veteran system design interviewer just dropped a spicy rant: stop reading answers off ChatGPT, do your homework on the company (he even name-checks Rokt), start with the customer, and it’s okay to say “I don’t know” (he says that’s a plus). His pet peeve? The glasses glare that betrays candidates tabbing between AI and the shared whiteboard. He’s begging for your real voice, not the “global leader in buzzword bingo” script.
But the comments turned it into a cage match. One commenter says the advice is fine but basic, adding that “making money isn’t an engineering problem” and bragging their company 5x’d revenue with a simple Rails app plus killer marketing and tiny UX tweaks. Another blasts the post as one‑sided, claiming they were rejected for not using enough AI—yes, some interviewers want you consulting a bot every 10 minutes. The crowd split into camps: Team “do your research and be human” versus Team “these interviews are broken and signals are chaos.” Memes flew about “AI doping tests,” “Excalidraw tab-switch triathlons,” and the dreaded buzzword-laced company intros. The mood? Entertained, annoyed, and very unsure what interviewers actually want.
Key Points
- •Barker evaluates candidates on knowledge depth, independence, collaboration, prioritization, purpose, and decision-making under uncertainty.
- •He warns against overreliance on AI tools like ChatGPT during interviews and prefers authentic, candidate-driven reasoning.
- •Admitting “I don’t know, yet” and asking for help is presented as a positive signal of self-awareness.
- •Candidates should research the company’s products and economic model ahead of interviews, aligning with Warren Buffett’s advice.
- •Effective system design starts with customer needs; use structured questions (e.g., Five Ws) to clarify the user and purpose.