Comet Plus Launches AI-Human News Platform
Today's curated insights on AI product management from X/Twitter across 60+ expert sources and YouTube channels.
Comet Plus Launches AI-Human News Platform
From X
AI Product Launches & Updates
Comet Plus launch: Arav Srinivas @AravSrinivas announced Comet Plus, positioning it as the “Apple News + for AIs and humans” to consume internet content.
Agentic v0 update: v0 Team @v0 shared that they’ve shifted from “never touching tools” to letting them “run free,” unlocking more autonomous agent behaviors.
System instructions editor: Logan Kilpatrick @OfficialLoganK revealed a new System Instructions button in the interface, enabling full editing and reordering of message chunks.
AI Tools & Applications
AI agent tech stack: Aakash Gupta @aakashg0 listed all the top AI agent tools in one comprehensive thread to help PMs evaluate and integrate them.
Automated revenue workflows: Aakash Gupta @aakashg0 explained how he used Zapier + Claude to scan emails every 2 minutes, summarize them, and draft replies—generating over $1 M in 12 months.
Podcast processing agent: Claire Vo @clairevo showcased a custom-built “Parakeet Podcast Processor,” a terminal-based agent that converts 36 hrs of audio into market insights and blog content.
Product Management Insights & Strategies
MBA training critique: Shreyas Doshi @shreyas argued that traditional MBA curricula (case studies, participation grades) can hinder real-world product success, offering a 5-minute video breakdown.
AI PM fast-track: Aakash Gupta @aakashg0 distilled the future of AI Product Management into an 81-minute session from OpenAI’s Head of Integrity Product, ideal for PMs upgrading their skills.
AI Industry Developments & News
AI knowledge ceiling: Lenny Rachitsky @lennysan highlighted how human experts must produce fresh training data—like correcting protein folding steps—to keep models (ChatGPT, Claude) improving.
AI PM hiring boom: Aakash Gupta @aakashg0 noted that companies (even Apple) are hiring AI PMs, offering 30%+ salary premiums and signaling strong market demand.
Reasoning AI agents: NVIDIA AI @NVIDIAAI kicked off a thread on how reasoning agents can think through problems, toggle reasoning to optimize compute, and drive decision-making in medicine and finance.
From YouTube
How to digest 36 weekly podcasts without spending 36 hours listening | Tomasz Tunguz
How I AI Podcast • August 25, 2025
Tomasz Tunguz demonstrates his Parakeet Podcast Processor, a terminal-based AI pipeline that automatically downloads, transcribes, and cleans 36 weekly podcast episodes, generates daily summaries with key themes, quotes, tweet drafts and startup mentions, and then auto-drafts and self-grades blog posts using an AP English teacher prompt loop.
Key Takeaways:
- The Parakeet Podcast Processor fetches RSS feeds for 36 podcasts each day, uses ffmpeg plus OpenAI Whisper or Nvidia Parakeet to transcribe audio, and cleans transcripts via Gemini-3 through Olama locally.
- Daily AI summaries include host and guest details, comprehensive summaries, key topics, themes, notable quotes that feed venture investment theses, tweet suggestions, and extracted company mentions for CRM enrichment.
- A Python blog post generator pulls from a vector DB of Tomasz’s 2,000 articles to match his style, drafts a post, then iteratively grades and refines it three times with an “AP English teacher” prompt until it reaches an A- quality before final editing.