February 6, 2026
Scissors vs Servers
Why I Joined OpenAI
From haircut epiphany to OpenAI: mission vs money showdown
TLDR: A top engineer joined OpenAI to speed up ChatGPT and cut data center costs, pitching it as planet-saving. The community split fast: some say it’s noble mission, others say it’s about cash and an IPO, with privacy hardliners demanding no user data be read or used.
A star performance engineer says he joined OpenAI to make ChatGPT faster and cheaper—and “save the planet.” Cue the internet: one camp is cheering the move as a big-brain hero taking on big-scale problems; another is yelling “just say you joined for the money.” The spark? His heartfelt tale about a hairstylist who uses ChatGPT daily and loves its memory feature. Some readers got fuzzy feelings; others called it sad that a bot now fills in for human connection.
The hottest take: AGI—normally “artificial general intelligence”—got rebranded to “A Glorious IPO.” The “save the planet” line drew eye-rolls and eco-skepticism, while privacy hawks demanded the new hire ensure prompts are never read or used for training, ever. Meanwhile, fans hyped him up: he had 26 interviews with AI giants, and one commenter said he’s so good he shouldn’t have needed any—now the pressure’s on OpenAI not to mess it up.
In short, it’s a classic internet split: optimists love the mission, cynics smell stock options, and the privacy crowd wants ironclad guardrails. The comments turned a career move into a mini culture war—scissors vs servers, planet vs profits, bot comfort vs human connection—and everyone brought popcorn.
Key Points
- •The author has joined OpenAI to focus on performance engineering for ChatGPT and AI datacenters.
- •They argue that traditional performance engineering is insufficient at current AI scale and aim to develop new methods for larger, faster optimizations.
- •Anecdotal evidence from non-technical users (e.g., hairstylist) convinced the author of ChatGPT’s broad, practical adoption, including its memory feature.
- •Past experiences at Netflix and Intel informed the desire to work on a widely recognized product with large-scale impact.
- •The author held 26 interviews with AI tech companies and concluded that optimization challenges span the entire stack, not just GPUs.