Hey all,
After a year of skepticism, I finally tried LLM-assisted coding, and I'm genuinely impressed! I can now spin up multiple instances of Xcode and VSCode and churn out code at a speed I never thought possible. Yes, the majority of the time the code needs significant refinement and guidance, often feeling like I'm dealing with an unmotivated teen. However, with enough focused prompting, it can produce decent results.
This prompt-review-merge loop, however, has made me the bottleneck in my own workflow. I wish I could outsource basic PR reviews to truly scale this process.
I also see a "gold rush" among non-engineers using LLMs for "vibe coding." While they're excited, they often overlook critical aspects like exposed API keys, SQL injection vulnerabilities, and ballooning cloud costs. They clearly need professional engineering oversight.
It became clear there's a missing piece for both scenarios: a human in the loop.
So, I decided to build a solution to bridge this gap and meet the demand for specialized support during this LLM-powered gold rush. I created vibeplace.ai.
vibeplace.ai is a double-sided marketplace connecting "vibe coders" with experienced professional engineers. Whether you're looking to take an idea from zero to one, scale from one to N, or simply resolve a technical bottleneck like my own, vibeplace.ai provides the expertise you need.
Do you see yourself on either side of this marketplace?
Vitalik
This is an interesting idea! I've found that the valley between customers and being able to work directly with a helper like you describe is too vast for meaningful interaction to work.
If the customer is paying enough, it can be worth it. But it would take a decent developer to be the helper in this situation, and that developer could likely make more money doing other work, so it makes it hard to grow this model in my mind. What are your thoughts on attracting talent to the provider side?