Ahmed AlNeaimy

The 2026 Setup Clean & Simple♥️ Let’s build Kasper Company Professionally.

I spent 10 hours from two of my Pro 20x GPT sub using Clawpatch v1 and v2 results are insane! GPT /review -> no findings! Clawpatch review -> 455 findings ( valid and real ) and still working @OpenAIDevs guys, it's time to see why default /review is not working perfectly! well done @steipete

Guys @OpenAIDevs @sama @thsottiaux i can confirm Codex is eating the usage limits so fast, there is something wrong with all my 4 Pro 20x accounts! I even didn't use them heavily cuz i was in a travel, back to work yesterday and today and runs out the entire week limits.

The brand new @antigravity version 2.0 Failed to login .. nice signal.

New Antigravity IDE version 2.0 New 3.5 Flash models "High and Low" No new Opus models there.

Antigravity models Pro weekly limits seems removed just like Ultra, and only 5-hours limits available which is good if this true!

Antigravity Gemini 3.5 Flash High is soooo fucking fast! i didn't expect this at all.

Antigravity Gemini 3.5 Flash High - running multi-agents finally!

Guys @antigravity WHEN THIS WILL END? it's really strange, more than 5 months same error all the time.

They resets all usage limits! thanks @OpenAIDevs you gonna make us not leave our Macs :P

WOOOOW !!! This is the new Gemini 3.5 Flash model, i swear to god i haven't even touched the code yet, only asked the agent for reviewing my codebase which is a very simple question, i never did anything else and the usage went from 100% to 40% in about 5 minutes or even less. Good luck with a new @antigravity drama, back to work like always with Codex.

More and more bugs, issues, problems .. etc @antigravity fix your products before you release something.

Gemini 3.5 Flash simply "BenchMaxed", if you don't know what BenchMaxed means, they optimized the model to get higher rank on benchmarks, that's mean the model has been pushed to the "Max" to show a higher results into the "Benchmarks", it doesn't have a real capabilities or high intelligence, it can't even finish a single fucking task without hallucinate, stuck, looping forever .. etc, this is what i actually saw with a real project, it doesn't even follow what i asked for at all! Always remember, any "Flash" model = full of shit, lower intelligence and capabilities are the trade off of speed. Companies are "Lying" with their benchmarks, never trust them, you can simply test the models with your REAL work and you can see the real results so you can judge by yourself instead of keep follow their "BenchShit".

Opus 4.7 Max with ~15 agents works well for over 1.5 day, refactoring a huge project, cleaning up dead-code, building a comprehensive e2e/units/integration testing The usage limits is really good for now, let's see the final results, i know myself i can't trust Opus code 100%, that's why i prefer GPT more.

Codex /goal still working amazingly over 13 hours non-stop, fixing and verifying 688 real bugs discovered using Clawpatch, 253 remains.

Gemini, Opus and everything is completely useless into @antigravity , man literally TWO prompts "review and plan" nothing else, and runs out of usage limits This is the Pro plan btw, i have Ultra plan and very sure the same shit there too, and it doesn't even using my AI Credit Overages at all.

@pradanaerza @antigravity Yes.

@AhmedAlNeaimy @antigravity got same problem, turn out antigravity load ALL GLOBAL SKILL in your prompt by default? or im missing something here lol

@_srimzzy_ See my tweets, I’m already a power Codex user.

@AhmedAlNeaimy Move to codex...

@GoogleAIStudio Fix your products then ask us to use them.

ai studio mobile: build your ideas, wherever you are coming soon to an app store near you ai.studio/mobile

Btw .. I lost 750$ of 3 Ultra plans in the past 3 months because @antigravity doesn’t work at all, and they didn’t even provide any compensation, they must be aware of their work.

@yvg_dev @OpenAIDevs @sama @thsottiaux I have a software development & ai company, that’s why I use a lot of subscriptions.

@AhmedAlNeaimy @OpenAIDevs @sama @thsottiaux Out of curiosity what are the complexity of tasks you’re carrying out, because I have a 5x GPT pro subscription and I’m having no issues with usage