AI passionate.
Outer Heaven
Joined July 2025
- Tweets753
- Following66
- Followers13
- Likes2.8K
I can see that
We vibe coded the Epic Games Launcher to be faster than the other guy.
Don't believe us? Test it out for yourself in Preview:
launcher.epicgames.com
Google seems to be really afraid that people will forget about it's unreleased, delayed to never, new Pro model.
This quoted post is unavailable.
Antigravity taking minutes to make a single edit.
This has been happening intermittently over the past few days.
AI studio competing against GPC for the title of worst UX:
1. added a billing account on GDC, paid a few dollars.
2. get to ai studio, asked for another 10 dollars fee. paid.
3. another window, asking for a pre paid entry fee (again).
I don't understand the Gemini app.
Once, I asked it to create a document to help me keep track of something. It actually connected to Google Drive and created one there, which was unexpected.
Now, when I ask it to create one there in a new chat, it just says it can't.
Peak UX.
I never have heard about this company, and the first time I see it, it's straight up lying by omission, selling paid features saying it's free.
Congratulations.
Introducing Bolt Forge. Free until Oct 14th:
- Up to 50x more usage
- The new frontier: GLM, DeepSeek, Kimi
- Zero usage charges
Live now in your model picker on bolt.new
And one more thing... 👇
Gemini app is messed up. I have a chat that every messages trigger a Imagem generation.
And I mean it.
My quota was nuked.
Since you guys don't looks like fixing it, can we please just have a switch to completely disabled this feature?
@GeminiApp @joshwoodward
ALT It generates the normal text fallowed by an image. Nothing on my input had any hint to trigger this. Then I asked "what? I didn't ask for any image" and it just generated another one.
ALT "To confirm that the text output instruction has been assimilated and that no extra visual elements will be spontaneously generated, this technical image illustrates the blockage." Literally joking with me.
I'm more interested in Flash models.
Pro is great, but if we keep having those Flash iterations with this level of jump in intelligence, most of the use cases will become trivial.
That is, if they keep the price as it is right now. That means token efficiency too.
Gemini refuses to answer me about this image with guitar tabs.
it starts to generate text, but its cut mid through and replaced with "i cant help with that".
@GeminiApp @joshwoodward
I have had really bad experiences with Gemini spark so far.
1. Undercooked. The agent errors out a lot.
2. Painful slow for even basics.
3. I can't queue messages in the chat.(?)
4. It's way dumber than any Gemini model, and too slow to be flash lite.
Gemini App is getting worst.
Apparently, we don't have code execution anymore. This is very frustrating. I don't want to use spark for a quick thing.
Could you let me use my app as I want, please?
Ive just lost minutes trying to use blender mcp with antigravity. something that works out of the box with other harness.
I try to be positive, but to every Gemini 3.7 there's other 20 bs like this breaking the UX.
Its still not working, btw.
Meanwhile, Gemini App is moving backwards on Android. Now it doesn't open quizzes the model generates natively. It has to open a webpage to show it.
At this point, saving the webpage as pwa is better than the native App.