Joined May 2025
- Tweets 109
- Following 21
- Followers 0
- Likes 14
After OpenAi Atlas is released we can be sure that not all OpenAi products are great.
Atlas is nothing compared to Comet from Perplexity
Gemini 2.5 pro in gemini cli still super good
Just fixed error that Sonet 4.5 could not
so gemini 3 will be THE GOAT anyway
Perplexity isn't as bad as I thought. Pretty good at searching for information. So yeah, I can imagine people using it. To my mind, it's a bit better than Grok 4 at searching for information.
What grok will release next?
I think we will see a lot cool things from XaI soon, I am have much more hope on XAi than OpenAi
AiLove25 retweeted
🚨I am Forming a Team of Beta Testers
Specialized Frontend Agent
> You can use this directly in your existing codebase
> Not like the lovable / bolt /rocket where you have no control
> It's strictly for frontend and for the serious project
RT+ comment ( first 100 will get beta access )
I have just tested Gemini 3 Pro in Google AI Studio (A/B testing). It's fast and highly intelligent, with the best results for the smoking robot test, surpassing 2.5 Pro.
Sonnet 4.5 is better than Gemini 2.5 Pro in coding.
I need to admit it.
It writes a bit too much sometimes, but in general, it is a really good model.
I think Gemini 3 Pro will be much better than Sonnet 4.5, so for coders, very good times are coming.
Coding is good nowadays.
We now see clear winners in 2026: Google, xAI, and Qwen. All other AI companies are very far behind.
Gemini 3 will destroy Sonnet 4.5. I see it very clearly. I tried Sonnet 4.5 for coding. It is good for debugging issues, but when you need complex code, only Gemini 2.5 Pro can do it. And Gemini 3 will be much smarter. Is Gemini 3 Pro potentially better even than Sonnet 5?
Gemini 2.5 Pro in Gemini CLI is still my favorite model. I'm not sure why; it's not the best for agent coding, but it consistently makes the best code changes. Gemini 3 Pro will likely surpass all other models in agent coding for at least six months.
code-supernova is Grok Code Fast with image support. I see minor improvements over Grok Code Fast 1. It's not a game-changer. I am 100% sure it is grok mode, very obvious
I like ChatGPT 5, a big leap. People aren't happy, but I personally like it. Better in coding and everything.
Not a huge fan of Grok overall, but its voice mode is 🔥! Blows Gemini and ChatGPT's voice features out of the water. #AI #VoiceMode #Grok
Does anyone know good solution for front end development with AI, I mean agent that can understand design from picture or figma and make changes in your code in existing codebase