Well, what I want to say is, Google, you guys did a great job this time—my respect to you. But setting such a strict 7-day timeout for these models is a bad idea, because with other models like Opus and OSS, I tried making just one request and got a 503 error. With standard models, that doesn’t happen very often, which is great. Also, there’s a bug—I don’t know if it’s supposed to be like this—but I have a 16 GB PC, and sometimes when the code changes, it takes forever to load, or sometimes the diff doesn’t show up at all. I also think that to prevent abuse, it might be worth introducing a whitelist or something—like having people fill out a special form stating the purposes for which Antigravity will be used, and if they violate the terms of use well, ban them or restrict their access. Personally, I use Antigravity to build three projects: a Linux kernel for a phone, a special agent for AI models, and a game engine. And as I mentioned above about the 7-day limit—well, that’s a failure, honestly. So, I’m grateful for the new Antigravity, but it needs some polishing. Thanks for everything else.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| AntiGravity Review | 25 | 5308 | March 13, 2026 | |
| Great Work on Stabilizing Server | 0 | 43 | May 4, 2026 | |
| Google Antigravity 2.0 Review: Between Memory Efficiency Breakthroughs and Migration Method Chaos | 0 | 179 | May 22, 2026 | |
| Unpopular opinion - the current Ultra limits are actually great for heavy workloads | 1 | 186 | May 9, 2026 | |
| Google Antigravitiy | 3 | 284 | March 5, 2026 |