Let me be perfectly clear: I love ChatGPT the way a Wall Street exec loves Adderall—functionally, obsessively, and with a creeping sense of dread. And yet, here I am, sitting slack-jawed, waiting for my beloved GPT-4 to respond like it’s typing on a Nokia from 2007. In those 20-second limbos, I find myself not pondering the great questions of our time, but wondering, “Did Sam Altman outsource the servers to a hamster wheel in Palo Alto?”
Yes, this is a response to the brilliant piece “What To Do While I Wait for Claude”, and yes, I too have opened 37 browser tabs, changed my VS Code theme twice, and eaten four croissants before ChatGPT gave me a code snippet with a typo in it. But I come bearing a different energy. Less “frustrated engineer,” more “investment guy who discovered GPT and ended up firing a lot of staff.”
First, A Confession: I Replaced Many Staff Members with OpenAI This Year
It started small. First, the SDRs.
We had a B2B sales team. Five humans. Great people, okay results. Enter ChatGPT fine-tuned with HubSpot data, a Notion CRM export, and a couple sketchy browser plugins. Suddenly, follow-up emails became automated, demo scheduling ran 24/7, and our lead conversion rate… went up. The humans? Let go, kindly. One started a candle business. (Respect.)
Then came the Jr. Software Developers. Let’s be honest—a shocking amount of junior dev work is Googling and gluing API docs together. ChatGPT does that faster and doesn’t ask if we have dental. A prompt, a spec, and suddenly I had a fully functioning Firebase backend. I even had it auto-document everything in Markdown and push to GitHub via OpenAI’s code interpreter. The junior dev team now includes me, ChatGPT, and an espresso machine.
Next? Recruiting. We used to pay $12k per hire to agencies. I fed job specs and our ideal candidate profile into GPT and piped it into PhantomBuster and LinkedIn automation flows. We made 17 hires in 3 months. All GPT-sourced. No recruiter fees. No awkward Zoom calls. One candidate asked if our recruiter was AI. I lied.
Marketing? Rewritten in GPT-4 using tone presets. HR onboarding? Streamlined using GPT-generated Notion docs and Loom videos. Intern tasks? Simulated and auto-graded via GPT + Typeform. QA? LLM-assisted test-case generation through LangChain. Legal doc reviews? GPT with a twist of CaseText.
I didn’t set out to become a capitalist AI villain. But when you get 30x leverage from a $20/month subscription (plus a couple GPT Store hacks), it’s hard not to feel like Tony Stark with a moral blind spot.
ChatGPT: The Master of Suspense
But here’s the twist. Just like Claude, ChatGPT is also a master of suspense. You ask it to debug a regex, and it stalls like it’s waiting for approval from legal. Or worse—you get the dreaded “Hmm… something went wrong.” Or my personal favorite: “ChatGPT is at capacity.”
Oh, you’re at capacity, are you? I just fired Steve and Priya because you said you could handle their workload, and now you’re telling me you’re too tired to answer my prompt?
Is Sam Altman overclocking GPT’s servers while feeding them a diet of investor decks and GPT Store widgets? Possibly. Is ChatGPT increasingly acting like a 10x engineer who knows he’s irreplaceable and therefore works on his own time? Definitely.
Sam Altman, Respectfully: Your Product is Becoming a Moody Teenager
Look, Sam’s done a lot. The Worldcoin iris-orb thing—psychotic. The OpenAI boardroom coup saga—Pulitzer-worthy. The wardrobe? Predictably monotone. But let’s talk about ChatGPT—your golden child. It’s developing some unsettling habits:
- Takes forever to respond when you need it most
- Gets weirdly evasive when you ask about OpenAI governance
- Drops context like a crypto project drops ethics
GPT-4 is like if Sherlock Holmes was brilliant, helpful, and occasionally wandered into traffic. I still rely on it for everything from summarizing lawsuits to refactoring 2012-era Ruby scripts, but it needs a nap and a new GPU cluster.
So What Do I Do While I Wait?
- Replace another department with AI
- Watch Alex Hormozi reels until I want to sell my pancreas for a higher ROAS
- Prompt GPT to write a script about how slow GPT has become
- Rewrite my entire prompt 3 times to make it “easier for GPT to understand”
- Drink espresso
- Walk to a further away bathroom on my bathroom break
- Blog on HackerNoon
- Start an internal monologue about how AI could in fact do my job
The Inevitable Future – ChatGPT Will Load Again!
AI isn’t coming for your job. It already came, said “nice resume,” and then built your replacement. While you’re waiting for ChatGPT to finish loading, it’s automating another slice of your workflow in the background.
So yes, I wait for ChatGPT. But I do it while laying off another internal process, updating my job title to “AI-native Ops Strategist,” and considering whether I need any employees at all.
Sam, if you’re reading this, fix the latency. Or at least give me a loading animation with a mustache.
Yours in full automation, An Anonymous AI Operator
