AI SDK 5 Now Supports Google Gemini 2.5
Today's curated insights on AI product management from X/Twitter across 60+ expert sources and YouTube channels.
AI SDK 5 Now Supports Google Gemini 2.5
From X
AI Product Launches & Updates
-
Video Generation with Audio: Aravind Srinivas @AravSrinivas announced that video generation (with audio) is now available on Perplexity for Pro and Max subscribers, requiring mobile app updates and offering higher rate limits for Max users.
-
Google Gemini 2.5 Support in AI SDK 5: Philipp Schmid @_philschmid shared that AI SDK 5 now supports the full Google Gemini 2.5 model family with a configurable "thinking config" and built-in tools like Google Search.
AI Tools & Applications
-
Internal AI Assistant “Moda”: Claire Vo @clairevo highlighted How I AI Episode 18, showcasing how @Amplitude_HQ built Moda, an internal assistant that unifies access to company data across multiple systems.
-
ChatGPT for Product Managers: Aakash Gupta @aakashg0 shared AI 401: AI Building, a course on leveraging ChatGPT to prototype and build AI-powered products.
Product Management Insights & Strategies
-
Rapid ARR Growth: Karan Vaidya @KaranVaidya6 recapped a ComposioHQ x AI Plus panel where a founder explained hitting $35 M ARR in 45 days through maniacal data obsession and reckless execution speed.
-
Shipping Fast to Discover Use Cases: Lenny Rachitsky @lennysan noted that OpenAI’s Head of ChatGPT advises to ship quickly to uncover real AI use cases before polishing features.
-
Engaging Engineers in Discovery: Teresa Torres @ttorres recommended practical strategies to link product discovery to decision-making, ensuring engineers stay invested in user-centered innovation.
AI Industry Developments & News
-
User Attachment to Models: Sam Altman @sama observed that the emotional attachment to AI models like GPT-5 is unprecedented, marking a notable shift in how people relate to AI.
-
Open Source in AI: Clement Delangue @ClementDelangue highlighted the critical role of open source in AI, pointing to global competition and commending both OpenAI and Meta for contributing to open-source efforts.
From YouTube
How to Build Your Entire AI Workforce in One Afternoon (Live Demo)
Greg Isenberg • August 11, 2025
Greg Isenberg and Flo Crivello demonstrate Lindy AI’s new Agent Builder and Computer Use features, showing how anyone can build, iterate, and deploy AI agents in plain English to automate tasks like LinkedIn outreach, customer support, and recruiting in under an afternoon.
Key Takeaways:
- Lindy AI’s Agent Builder and Computer Use address the two main AI agent hurdles—complexity for non-engineers and integration with existing tools—by letting users describe workflows in natural language and automating any desktop action.
- In the live demo, Greg and Flo built a LinkedIn outreach agent in under three minutes: it triggers on LinkedIn profile URLs, logs into LinkedIn via Computer Use, and sends personalized DMs.
- Agent swarms can parallelize workloads: a 20-agent swarm searched for US-based software engineers at Zapier, deduplicated contacts, and sent follow-up emails within three minutes using Lindy AI.
How Amplitude built an internal AI tool that the whole company’s obsessed with (and how you can too)
How I AI Podcast • August 11, 2025
Wade Chambers, Amplitude’s Chief Engineering Officer, explains how his team built Moda—an AI-powered Slack bot and web interface that aggregates over a dozen enterprise data sources using Glean and a custom “Langley” framework in just a few weeks to enable companywide search, PRD generation, and rapid prototyping.
Key Takeaways:
- Moda indexes internal systems like Confluence, Jira, Salesforce, Zendesk, Slack, Google Drive, ProductBoard, Zoom transcripts, Outreach, Gmail, Asana, Dropbox, GitHub, and HubSpot via Glean APIs for on-demand AI queries.
- The team delivered a Slack-native bot and a proprietary web UI in 3–4 weeks of engineers’ spare time, orchestrated by Amplitude’s Langley framework to power multi-step AI workflows including PRD creation and prototype prompts.
- Amplitude drove viral adoption by making Moda usage public in Slack channels, using social proof to reach companywide engagement within a week and iterating via GitHub-managed YAML prompt configurations.