The Future of AI Studio and Gemini ( Logan Kilpatrick @ Google Cloud Next)

Sam Witteveen · Beginner ·🧠 Large Language Models ·5h ago
In this video, Sam sits down with Logan Kilpatrick to talk about the latest across Google's AI ecosystem — from AI Studio updates and TPU infrastructure to Nano Banana 2, Antigravity, and more. Twitter: https://x.com/Sam_Witteveen 🕵️ Interested in building LLM Agents? Fill out the form below Building LLM Agents Form: https://drp.li/dIMes 👨‍💻Github: https://github.com/samwit/llm-tutorials ⏱️Time Stamps:00:00 - Intro 02:02 - AI Studio overview — from prompt to prototype to production 02:22 - The Build tab — full vibe coding experience 04:06 - New: Design previews 04:25 - New: "I'm Feeling Lucky" button 05:04 - New: Tap tap tab — AI prompt autocomplete 06:20 - Opinionated stack — Next.js, Firebase, Cloud Run 09:16 - Audio models: TTS, Live, Lyria 11:00 - Mobile AI Studio 11:55 - Vibe coding → agentic engineering 15:00 - The "ambition" mindset shift 21:00 - AI Studio for Everyone 21:23 - Gemini 3 / 3.1 model discussion 22:25 - Gemini Live — real-time audio + video streaming 24:40 - Project Astra — precursor to Gemini Live 25:41 - TTS model adoption + function calling via voice 27:14 - Nano Banana + Gen Media portfolio 29:07 - Coding models 29:37 - Antigravity 31:19 - TPU announcement 33:55 - What's next after coding? 35:32 - Deep Research API - New Launch 43:02 - Wrap up
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Related AI Lessons

⚡
Ein Echo hat keinen Mund.
Understand how language models work and why they don't have a sense of self, to avoid anthropomorphizing them
Medium · AI
 ⚡
The Discipline Is the Product
Apply four lenses to build working AI systems by identifying where judgment belongs, stripping unnecessary tasks from LLMs, and ensuring the system meets actual needs.
Medium · Machine Learning
 ⚡
I benchmarked 10 LLMs on slopsquatting — up to 87% installed fake packages
Benchmarking 10 LLMs on slopsquatting reveals up to 87% installed fake packages, highlighting security concerns in AI package management
Dev.to AI
 ⚡
ChatGPT Image 2.0 Signals Visual Reasoning To Solve Real-World Tasks
ChatGPT Image 2.0 enables visual reasoning to solve real-world tasks, marking a significant advancement in AI capabilities
Forbes Innovation

Chapters (22)

2:02 AI Studio overview — from prompt to prototype to production
2:22 The Build tab — full vibe coding experience
4:06 New: Design previews
4:25 New: "I'm Feeling Lucky" button
5:04 New: Tap tap tab — AI prompt autocomplete
6:20 Opinionated stack — Next.js, Firebase, Cloud Run
9:16 Audio models: TTS, Live, Lyria
11:00 Mobile AI Studio
11:55 Vibe coding → agentic engineering
15:00 The "ambition" mindset shift
21:00 AI Studio for Everyone
21:23 Gemini 3 / 3.1 model discussion
22:25 Gemini Live — real-time audio + video streaming
24:40 Project Astra — precursor to Gemini Live
25:41 TTS model adoption + function calling via voice
27:14 Nano Banana + Gen Media portfolio
29:07 Coding models
29:37 Antigravity
31:19 TPU announcement
33:55 What's next after coding?
35:32 Deep Research API - New Launch
43:02 Wrap up
Up next
5 Levels of AI Agents - From Simple LLM Calls to Multi-Agent Systems
Dave Ebbelaar (LLM Eng)
Watch →