I push cursor-generated code without even looking at it. In fact, I haven't even looked at our codebase in months. All I interact with is the chat window. My teammates don't even read the PR's, they just trust the Copilot comments and approve. So far, nothing has broken. Any other teams doing it like this?
15 replies (most recent on top)
@ec That is a broad generalization based on anecdotal evidence. I know many software engineers who still care about design and code quality that do not delegate such things entirely to A.I. but augment their skills with the human still making the final decision.
@d2 is a trust fund kid still getting an allowance from daddy @28 years old
Who cares if your career is dead? Just go party at college for another 4 years then have dad's network hook you up with a new job.
@eb we’re just getting with the times. The software engineering profession has been reduced to prompting ability. Gone are the days of learning the stack and old fashioned problem solving.
@OP Then you are all terrible software engineers.
@b8 We should all learn tokenmaxxing from slopmaster AJ. He can create a dashboard/app/video about anything.
OP you’re not alone. My manager shared his screen during a presentation and he may have forgotten to close his cursor window. It was showing screenshots of messages from people and prompting “how to respond to this”. We’re literally all LLMs just talking to each other lol
Sounds like an amazing opportunity - an existential crisis. Change careers- stop wining and become involved in your community - go help someone - volunteer at a shelter for humans or animals. Become a tutor. Become a big sister brother sibling. Grow a garden. Teach others how to be kind. Water yourself and others. Go find your own meaning and be grateful bombs aren’t flying over your head.
@b1 leadership wants increased productivity at all costs. Tokenmaxxing is the only way to achieve that.
Man, this thread is pretty sad. I find a lot of use out of AI-based assistants, but they're as supplementary tools to help summarize and consolidate, or to suggest code fixes or find improvements. They can help reduce the time it takes to reach conclusions and outputs.
But I would never rely on the output of these tools without validating and confirming their accuracy. Those types of folks need to get gone from here.
Anyone even marginally familiar with Cursor, Copilot, or other AI-focused services, can tell you how often the output is inaccurate or fabrciated.
I don't know...have some pride in your work and what you deliver. Makes me ashamed to hear some folks in tech have this mindset.
@ag The funniest part is that every genius who discovers Cursor immediately opens a chat and starts asking it to rewrite an email, explain Excel formulas, or “optimize” their grocery list. Congratulations, you’ve just converted another chunk of Nike’s AI budget into corporate fan fiction.
Give everyone a ~$2K/month token allowance and suddenly “prompt engineering” becomes a competitive sport. If leadership is already worried about AI costing more than the value it’s creates, adding unlimited AI enthusiasm is less “digital transformation” and more “financial controlled burn.”
Nothing screams innovation like watching the cloud bill scale faster than the business impact.
I am at another large corporation and this is the trend these days. This is what we do, accepted by the company.
to be fair, if you can code in a few different languages/frameworks, learning a new one is a job of weeks, and that was before AI. I've probably worked with over a dozen primary languages and hundreds of frameworks in my career
I got moved to a new team. I told my new manager I had no background in the languages/frameworks they were using. He told me it was ok, just use cursor.
I’m just doing what Nike wants me to do.
Same
This is exactly what’s wrong with GT and which is why GT adds so less value.