@orendev_ Actually I don't agree. Cursor always write the comments very well and never use variable names like"a1","b2". Even if there are more bugs delivered they are much easier to fix.
Every Gemini 4 model ships with a native 1M context window, yet @antigravity hard-caps it at 256k—even on Google AI Ultra. Crippling your biggest differentiator inside your own flagship IDE is an absolute shame. Unlock the full 1M! @OfficialLoganK@_mohansolo@kevinhou22@ivanleomk
appreciate the reply! Three specific points on this:
1. Totally understand 1M burns quota faster—just give power users an opt-in toggle (with a warning if needed) and let us make that tradeoff ourselves :)
2. 256k is totally fine when a build goes smoothly on the first pass. The issue hits when the initial deliverable misses expectations and requires heavy back-and-forth iteration—multi-turn debugging blows past 256k fast, and context compaction starts dropping critical constraints.
3. For long-form novel writing in the workspace, 256k vs. 1M is night and day—keeping full manuscript + lore consistency simply needs that 1M window.
Every Gemini 4 model ships with a native 1M context window, yet @antigravity hard-caps it at 256k—even on Google AI Ultra. Crippling your biggest differentiator inside your own flagship IDE is an absolute shame. Unlock the full 1M! @OfficialLoganK@_mohansolo@kevinhou22@ivanleomk