I wanted to get everyone’s thoughts on the current state of AI ecosystems—specifically open-source and Chinese platforms versus Google’s approach to “vibe coding.”
Honestly, I’m leaning heavily in favor of the open-source and Chinese models right now. Despite being a massive tech giant, Google seems to be rushing out of fear of falling behind, and it’s resulting in huge mistakes. If you check the Google Devs forum right now, the situation with Antigravity is a perfect example of this panic. They forced the 2.0 update, completely gutted the integrated IDE workflow, and introduced absurd bugs. We’ve got users hitting weekly Gemini 3.5 Flash lockouts after 30 minutes, the CLI (agy) failing to persist auth states, and quota refresh times glitching out to 6 days instead of 5 hours.
When you’re trying to push through a final-year engineering project—like the IoT-based air quality monitor I’m currently building—you can’t afford these kinds of roadblocks. Vibe coding is supposed to make development seamless. Whether I’m spinning up a React and FastAPI stack, or tweaking GLSL shaders for data visualization, I just need the AI to reliably assist. Trying to build clean, premium, and polished interfaces is incredibly frustrating when the underlying ecosystem drags you back to the stone age and arbitrarily locks you out of your own workflow.
Meanwhile, open-source communities and Chinese AI ecosystems are iterating fast and delivering highly capable, usable tools. They aren’t paralyzed by the same corporate panic, and they aren’t pulling the rug out from under developers with broken, untested forced updates.
Are we seeing a real shift where open-source becomes the default for actual development, while the giant ecosystems get bogged down? Is anyone else jumping ship to open-source alternatives because of these ecosystem bugs?
Would love to hear what you all think.