How much better will vibecoding get
openai.comHere's how I see it : Every company and startup will see vibe coding tools as the new get-rich-quick scheme, and then it will decline in quality and AI will hallucinate no matter what. So, yeah, for now, it will get better, but not for long.
Most interesting response to this, especially given all the "AI is going to replace software engineers by next year". My question would then be why would it stagnate if Cursor is just getting better and better with the new models? I mean for me it would seem it would slowly improve with time.
Yes, but the thing is, AI will become more and more expensive to run and maintain and build, and as every company/startup is going to the AI hype train. It would be harder and harder to meet the level-up demands for AI (newer and more powerful models), and then, every company will reach a point where every AI model, is gonna be pretty much the same, but getting more and more expensive to maintain, no company would create something new since they have probably reached the limit of their creativity and their resources (This is a vibe coding scope tho)
yeah makes sense