Google Gemini 2.0

@mihwa.jaeseung

11.9K followers · usually 2.8K views a post

Views

11.1K

Likes

227

Comments

5

Shares

66

SoundTikTok sound · original sound - mihwa.jaeseung

Promoted product

Google Gemini

Why it worked

The video effectively showcases a cutting-edge AI feature (Gemini 2.0's multimodal capabilities) in a practical, relatable scenario (coding assistance). The demonstration of stacking LLMs provides a novel and intriguing use case, appealing to tech-savvy audiences.

Summary

The video introduces Google Gemini 2.0's multimodal AI capabilities, highlighting its ability to process text, voice, and screen input. The creator demonstrates using Gemini within a coding environment to assist with prompting other LLMs, showcasing a seamless interaction that feels like having a developer assist.

Structure

1Introduction to Gemini 2.0's multimodal capabilities
2Demonstration of Gemini's ability to interact with screen content
3Using Gemini within a code editor environment
4Stacking LLMs for assistance
5Seamless and intuitive user experience

Product placement

Google Gemini 2.0: Acts - it is an AI model that can read text, hear voice, respond vocally, see the screen, and offer opinions. Removing it would change the video. Google Gemini 2.5: Acts - it is a text-to-speech model that can be used in a code environment. Removing it would change the video. WinSurf: Acts - it is a tool used to create a code environment. Removing it would change the video.

Original caption

Gemini 2.0 Voice Mode Feels Like the Future of AI #Gemini #GoogleAI #VoiceAI

More from @mihwa.jaeseung

On-screen text

Google Gemini 2.0 UNLIMITED AI VOICES Gemini 2.0 Voice Mode Feels Like the Future of AI July/31/26 TALK WITH GEMINI! Gemini 2.5 Text-to-speech Gemini 2.5 Flash Native Audio Function calling accuracy Adherence to developer instructions Overall conversational quality

Transcript

Google Gemini 2.0 today. We're gonna talk about it because it's a really big deal. This is now in multimodal AI. So this is one of those AIs that can read text, it can hear your voice, it can respond vocally, it can see your screen, it can read what's going on on your screen, it will have opinions about your screen. So what I did was I actually got a code environment up on my screen. I used WinSurf, you can use cursor, you can use whatever code environment you want. You don't have to have a code environment. You can make it a chat window, you can make it a word doc, you can make it an Excel spreadsheet. It'll work. And then I began to talk with Gemini. I said, you see my screen, I'm gonna talk with Gemini about it. In my case, I got Gemini to help me think through how to correctly prompt another LLM to code in the code editor. So I was effectively stacking LLMs, but it didn't feel precarious, it didn't feel weird. It actually felt like I had a developer over my shoulder and I was talking about what I was seeing in the code editor.