ByteDance's Seedance 2.0 and the latest in AI video, ads, and self-driving

Good morning, {{ first_name | AI enthusiasts }}. China's AI labs are racing ahead in AI video, and ByteDance's Seedance 2.0 may be the most impressive release yet. Viral demos from its beta show cinematic shots, consistent motion, and synced audio that look stronger than many current systems.

Seedance 2.0, ChatGPT Ads & Waymo World Model

Today's highlights

  • ByteDance's Seedance 2.0 stuns the AI video world
  • OpenAI begins showing ads in ChatGPT
  • How to build an AI-powered sales objection handler
  • Waymo uses Genie 3 to train self-driving cars
  • 4 new AI tools, community workflows, and more

ByteDance: Seedance 2.0 pushes the AI video frontier

The Rundown: ByteDance is going viral on Douyin and other social platforms with Seedance 2.0. In beta, the model produces smooth cinematic sequences, consistent character motion, and tightly synced audio. Early outputs suggest it may outpace many existing video-generation tools.

Key capabilities

  • Multi-modal input: text, image, audio, and video prompts reportedly work, producing strong outputs across styles from fight scenes to animation and user-generated content (UGC).
  • Native audio generation, 2K resolution, and 15-second clips available via ByteDance's Jimeng AI video platform.
  • Seedream 5.0 image model appeared in previews on third-party apps, positioned as ByteDance's answer to image models like Nano Banana Pro.
  • Released days after rival Kuaishou's Kling 3.0, signaling rapid advances in China's video AI scene.
Why it matters: Seedance 2.0 could spark a broad creative shift. Better cinematic coherence and audio sync make video more usable for creators and studios, and could disrupt production workflows.

OpenAI: ads arrive in ChatGPT

The Rundown: OpenAI started testing ads in ChatGPT for U.S. users on its free and $8/month tiers. Ads appear below chat responses and are targeted using the active conversation, chat history, memory, and past ad engagement.

Details and impact

  • OpenAI says ads won't alter ChatGPT's answers and aims to protect user trust for personal tasks.
  • Free-tier users can opt out of advertising, but doing so reduces their daily message allowance, nudging them toward paid plans.
  • Reported minimum ad spend for the pilot is $200K. Major agencies are already booking slots.
Why it matters: Monetizing AI through ads is a major industry test case. It preserves free access but could change user experience and the business dynamics of conversational AI.

AI training: build an objection-handler from sales transcripts

The Rundown: Turn your weekly sales-call transcripts into a practical AI reference for objections. This is a simple process that produces a reusable report for your sales team.

Step-by-step guide

  1. Create a ChatGPT project titled "Sales Objections." Upload a text file that lists products, pricing, and core offers so the model understands what you sell.
  2. In project settings → Instructions, paste: "Read the attached transcripts. Create a weekly report template. For every objection: Number the objection, state it clearly, provide 3 context bullets, list which lead surfaced it, provide two short rebuttals."
  3. Each week, upload call transcripts with client names and dates. Start a new thread and ask for the report.
  4. Open a Notion page and create a "Toggle Heading" for the week (e.g., "Week of Feb 9"). Paste the ChatGPT output inside the toggle.

Pro tip: Connect Notion inside ChatGPT (Settings → Connected Apps) so the model can reference past weeks and keep context consistent.

Waymo taps DeepMind's Genie 3 for self-driving simulations

The Rundown: Waymo introduced the Waymo World Model, a simulator built on Genie 3 that generates hyper-realistic driving scenarios the fleet hasn't seen. These synthetic edge cases help train cars for rare, dangerous situations.

How it works

  • Genie 3's visual knowledge converts into paired camera and lidar outputs to simulate realistic sensor data.
  • Engineers can edit scenes with text prompts or inputs—changing weather, adding obstacles, or reshaping layouts to test "what if" scenarios.
  • To overcome Genie 3's short memory, Waymo runs footage at 4x speed, stretching simulations to cover longer driving tasks.
Why it matters: World models let companies create extreme edge cases that real-world driving cannot reliably produce. This accelerates safer, more robust self-driving systems.

Trending AI tools and community updates

  • Codex App: OpenAI's macOS app for managing agents.
  • Composer 1.5: Cursor's updated agentic coding model.
  • Audiobooks: ElevenLabs' AI-powered narration suite for audiobook production.
  • Teleport: open-source blueprint for securing agentic AI; webinar on Feb 19 about identity for scaling agents.

Industry notes: Reports claim ChatGPT usage surpasses 10% monthly growth, Codex weekly usage rose 50%, and an updated model is expected soon. Anthropic may raise a large funding round, and ElevenLabs launched a full audiobook production suite.

Community workflow spotlight

Each newsletter highlights a reader using AI to work smarter. Today's submission comes from Clay D. in Carson City, NV, who used ChatGPT to analyze VA documents and successfully strengthen a disability claim—resulting in a higher rating.

Related coverage and gaming news

While this update focuses on AI video, ads, and robotics, gaming news continues to intersect with these advances. High-fidelity synthetic video and audio models can accelerate game cinematics, UGC content for streaming platforms, and procedural content generation. Expect studios and indie developers to adopt similar models for faster asset creation and better in-game narration.

See you soon,

Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown