OpenAI Launches ChatGPT Images 1.5
Today's curated insights on AI product management from 100+ sources across X, LinkedIn, and YouTube.
OpenAI Launches ChatGPT Images 1.5
From X
AI Product Launches & Updates
ChatGPT Images 1.5 launch: Sam Altman @sama announced Images 1.5 in ChatGPT and the API, offering faster generation, improved quality, and new editing capabilities.
SAM Audio release: AI at Meta @AIatMeta introduced SAM Audio, the first unified model for isolating sounds from complex audio mixtures, accompanied by benchmarks and research resources.
Claude Code enhancements: Claude @claudeai shipped updates for Claude Code, including syntax highlighting for diffs, prompt suggestions, a plugins marketplace, and shareable guest passes.
AI Tools & Applications
Notion integration for v0: v0 @v0 shared the ability to securely connect to NotionHQ workspaces, enabling teams to build tools and prototypes directly from existing docs and databases.
LlamaSplit beta API: Llama Index @llama_index introduced LlamaSplit, automating the separation of bundled documents into distinct sections to streamline mixed-document processing.
Vercel AI Gateway adoption: Guillermo Rauch @rauchg reported that Cline now uses Vercel AI Gateway, achieving 44% fewer errors and 14% faster streaming and auto-routing around LLM outages.
Product Management Insights & Strategies
Managing AI agent complexity: Lenny Rachitsky @lennysan recounted how Amir Klein described being overwhelmed by dispersed context when building an AI agent, highlighting the need for consolidated information channels across Slack, Notion, and more.
Building a PM second brain: Lenny Rachitsky @lennysan outlined steps to create a ChatGPT “PM second brain,” including defining its personality via prompts to boost productivity and knowledge management.
AI Industry Developments & News
$4B+ Databricks fundraise: Ali Ghodsi @alighodsi announced Databricks secured over $4 billion in funding, earmarked for Lakebase Postgres, Agent Bricks, and Databricks Apps.
AI agents survey results: LangChain AI @LangChainAI surveyed 1,300+ developers and business leaders to map agent use cases, production challenges, and tooling adoption.
Genesis Mission executive order: DeepLearningAI @DeepLearningAI reported a U.S. executive order launching the Genesis Mission, linking national labs, supercomputers, and private partners to train AI on federal datasets.
From YouTube
Testing ChatGPT's new Image model (full review)
Greg Isenberg • December 16, 2025
Greg Isenberg demonstrates how to access and use OpenAI’s new ChatGPT image model via the style carousel, showcases generating and editing images—such as plushie, sketch, hand-drawn diagram, and bobblehead styles—and compares its output quality and feedback capabilities against Nano Banana Pro.
Key Takeaways:
- Users can access the new image model by clicking the plus button in ChatGPT, selecting the image icon, and choosing from a style carousel featuring options like “3D glam doll,” “ornament,” “sketch,” and “plushies.”
- Using the “plushies” style and an optimized prompt, the model generated a highly detailed plush toy rendering of Sam Alman, capturing intricate hair waves and textures that exceeded Greg’s expectations.
- Greg observed that the model takes real-time instruction edits—such as removing notebook elements or adjusting sketch details—and delivers precise outputs, showcasing stronger instruction-following, creative transformations, and text rendering compared to Nano Banana Pro.
Shipmas Day 12: AI Music Video Generator App
All About AI • December 16, 2025
In this Shipmas Day 12 video, All About AI builds a UI-driven workflow that generates a complete AI music video—from Nanobanana Pro image and Gemini 3 lyrics to a Suno API-composed track and ByteDance OmniHuman (or Creative File Aurora) video—with automated chorus clipping and custom media upload features.
Key Takeaways:
- All About AI built a 12-step pipeline using Nanobanana Pro for images, Gemini 3 for lyrics, the Suno API for music, Whisper and FFmpeg for chorus extraction, and ByteDance OmniHuman to synthesize the music video.
- The app automatically transcribes the MP3 track with Whisper, identifies the chorus, and extracts a 10-second clip using FFmpeg for inclusion in the final video.
- A “bring your own clip” feature lets users upload custom images and audio files to bypass generation, and testing with the Creative File Aurora avatar model produced higher-quality video output.