OpenAI Expands ChatGPT Shared Projects
Today's curated insights on AI product management from X/Twitter across 60+ expert sources and YouTube channels.
OpenAI Expands ChatGPT Shared Projects
From X
AI Product Launches & Updates
Copilot humanist AI suite: Mustafa Suleyman @mustafasuleyman announced a set of updates—including Copilot Groups, an AI browser (live now), memory updates, and Copilot for health—and shared more details on the blog.
Shared Projects in ChatGPT: OpenAI @OpenAI expanded Shared Projects to Free, Plus, and Pro users, enabling collaborative chats, files, and instructions all in one workspace.
Major Google TPU expansion: Anthropic AI @AnthropicAI detailed a multi-billion-dollar expansion to secure one million Google TPUs and over a gigawatt of capacity for 2026.
AI Tools & Applications
Productivity AI toolkit: Theresa Iforit @theresanaiforit highlighted a suite of AI tools to 10× productivity—WisprFlow for 4× faster typing, VoiceNotes AI for voice memos, SciSpace for research, CodeRabbit AI for debugging, GuiddeCo for video guides, MeetGamma for slide creation, and more.
V0 AI app builder: V0 @v0 launched an AI SDK and gateway on Vercel, offering $5 in free monthly credits to spin up production-ready AI applications.
Annotate mode in Google AI Studio: Logan Kilpatrick @OfficialLoganK introduced a new Annotate mode that lets you draw UI changes directly on the interface and have Gemini apply edits into code seamlessly.
Product Management Insights & Strategies
10 steps to become a Google AI PM: Aakash Gupta @aakashg0 shared advice from a Google AI PM Director, stressing the importance of cultivating exceptional product taste and an intuition for 10Ă— improvements.
AI prompt library for experiments: George from đź•ąprodmgmt.world @nurijanian expanded his collection to 145+ mega-prompts, including an experiment design prompt that auto-generates 10+ hypotheses per feature concept.
Reality vs expectation in AI improvement: Lenny San @lennysan discussed insights from Chip Huyen on the gap between perceived tactics and what truly drives AI product success.
AI Industry Developments & News
AI safety through iterative refinement: Yann LeCun @ylecun compared AI development to turbojet reliability testing, arguing that real-world deployment and refinement are essential to prove safety.
Tech convergence era: Kevin Weil @kevinweil reflected on the simultaneous rise of powerful AI, quantum computing, and fusion energy, calling it an incredible time for innovation.
AI Dev 25 x NYC agenda: Andrew Ng @AndrewYNg unveiled the conference lineup, covering Agentic Architecture and orchestration frameworks from Google, AWS, Vercel, Mistral AI, SAP, and more.
From YouTube
Al Engineering 101 with Chip Huyen (Nvidia, Stanford, Netflix)
Lennys Podcast • October 23, 2025
Chip Huyen demystifies core AI engineering concepts—pre-training vs. fine-tuning, RLHF, and RAG—and shares lessons from her work at NVIDIA, Netflix, and Stanford on what factors actually boost AI product performance.
Key Takeaways:
- Debating the latest models, frameworks or vector databases yields minimal gains; the biggest improvements come from talking to users, building reliable platforms, preparing high-quality data, optimizing end-to-end workflows, and writing better prompts.
- Reinforcement Learning from Human Feedback (RLHF) uses human comparisons to train a reward model that guides the base model toward higher-quality outputs without requiring explicit labels for every example.
- Retrieval-Augmented Generation (RAG) performance depends more on data preparation—optimal chunk size, metadata, summaries or hypothetical Q&A—than on the specific vector database technology used.