ChatGPT Just Dropped and Everything Is About to Change

Five days ago, OpenAI released ChatGPT. I haven't stopped thinking about it since.

I was in my apartment in Boston, finishing up an assignment for my algorithms class, when my phone started buzzing. The Northeastern CS Slack channel was going off. Someone posted a screenshot of ChatGPT writing a recursive binary search with detailed comments. Someone else had it explain distributed consensus like you're five. A third person asked it to write a cover letter and the result was better than anything they'd written themselves.

I opened the tab, typed a question about transformer attention mechanisms, and got back a response that was clearer than most textbook explanations I've read. Then I asked it to write a Flask API endpoint with error handling, and it generated something I would have actually shipped. Then I just sat there for a while.

This Is GPT-3's Sequel, and It's Not Even Close

I wrote about GPT-3 back in 2020 when it first came out. I was impressed but skeptical. The API was powerful, but it was a developer tool. You needed to craft prompts, handle the API, build around its limitations. Normal people weren't going to use GPT-3.

ChatGPT is different because it's a product. You open a website, type in plain English, and get useful responses. No API key. No prompt engineering. No technical knowledge required. My mom could use this. My mom probably will use this.

That's the gap between a technology and a product, and OpenAI just closed it.

The Northeastern CS Slack Panic

Honestly, the conversations in our CS cohort have been wild. People are oscillating between excitement and existential dread, sometimes in the same sentence.

"This is incredible, also does my degree still matter?"

"I just used it to debug my homework, should I feel guilty?"

"If it can write code this well now, what does it do in two years?"

These aren't hypothetical questions anymore. I'm sitting in a master's program in computer science, surrounded by smart people who are genuinely unsure whether the skills they're learning will be relevant by the time they graduate. That's a weird feeling. Especially when you're paying American tuition on an international student budget.

What I Think Is Actually Happening

The thing is, I don't think ChatGPT replaces engineers. Not yet, probably not for a while. What it does is raise the floor. Tasks that used to require expertise, writing boilerplate code, explaining technical concepts, drafting documentation, are now accessible to anyone with a browser.

That's not a threat to good engineers. It's a threat to the parts of engineering that were tedious. And honestly, good riddance. I don't want to spend my time writing CRUD endpoints. I want to spend it on the hard problems. If ChatGPT handles the boring stuff, that's a win.

But it also means the bar moves up. If anyone can generate basic code, the value shifts to people who can architect systems, debug subtle issues, and make judgment calls about trade-offs. The skills that are hard to automate.

The Part That Scares Me a Little

I'll be honest. There's a part of me that watches ChatGPT generate a perfectly structured Python class and thinks, "I spent years learning to do that." The rational part of my brain knows that understanding why code works is different from generating code that works. But the emotional part? It needs a minute.

I talked to my friend about this over chai at our apartment. He's also in the CS program, also from India, also processing the same feelings. We agreed on this much: the people who learn to work with these tools, not against them, will be fine. The people who ignore them won't be.

It's been five days. The Slack channel is still going. I keep finding new things to try. Last night I had ChatGPT explain the CAP theorem using a Trader Joe's analogy and it actually worked. We live in strange and exciting times.