Seedance 2.0 Is Here… and It’s Better Than Sora & Veo - next-wave Recap
Podcast: next-wave
Published: 2026-02-17
Duration: 1 hr 4 min
Guests: Joe Fier
Summary
Seedance 2.0, ByteDance's new AI video model, offers groundbreaking improvements in video realism, challenging established players like Sora and Veo. The episode also covers significant advancements in AI tools, coding models, and the integration of ads in ChatGPT.
What Happened
The introduction of Seedance 2.0 by ByteDance marks a significant leap in AI video generation, enhancing realism through improved lip-syncing and consistent character portrayal. While currently accessible outside the US, global availability is anticipated by February 24th. Google is also making strides with Veo 3.1, which now allows advertisers to create video ads directly from text prompts within Google Ads.
Comparative analysis in the episode highlights Seedance's competitive edge over other models like Sora, Veo, and Runway, particularly in motion graphics and ad creation. The discussion underscores how AI is reshaping the creative industry, prompting a reevaluation of copyright and intellectual property norms, especially as US and Chinese models show differing leniencies.
The episode delves into OpenAI's GPT 5.3 Codex Spark, a cutting-edge coding model that can design websites in just 18 seconds. This model, optimized for speed with Cerebrus inference chips, represents a shift towards more efficient AI-driven coding solutions.
Google's Gemini DeepThink model emerges as a leader in scientific research, excelling in fields like chemistry and physics. It boasts a 50.5% score on the CMT benchmark and can convert sketches into 3D-printable files, showcasing its advanced capabilities.
The rollout of ads in ChatGPT, involving major companies like DoorDash and Adobe, signifies a new monetization strategy for OpenAI. Despite the introduction of ads, user trust is maintained through clear labeling, aiming to democratize access to AI on lower-tier plans.
The episode explores the broader implications of AI's rapid evolution, including the potential emergence of AGI through models capable of tool creation and use. Industries reliant on real-world skills remain less impacted by AI, highlighting the ongoing need for human creativity and expertise.
Key Insights
- Seedance 2.0 by ByteDance has pushed AI video generation forward with advanced lip-syncing and character consistency, making it a serious contender against models like Sora and Veo. The catch? It's only available outside the US until February 24th.
- OpenAI's GPT 5.3 Codex Spark can design a website in just 18 seconds, utilizing Cerebrus inference chips for speed. This represents a shift where AI is now not just a tool for coders, but a coding partner.
- Google's Gemini DeepThink model has scored 50.5% on the CMT benchmark, excelling in converting sketches to 3D-printable files. Its capabilities in chemistry and physics highlight AI's growing role in scientific research.
- Ads in ChatGPT are now a reality with companies like DoorDash and Adobe participating. Despite concerns about monetization, clear labeling aims to maintain user trust and expand AI access on lower-tier plans.
Key Questions Answered
What is Seedance 2.0 by ByteDance?
Seedance 2.0 is a new AI video model from ByteDance offering enhanced realism through improved lip-syncing and scene consistency. It's poised to challenge existing models like Sora and Veo in the video generation space.
How does GPT 5.3 Codex Spark improve coding efficiency?
GPT 5.3 Codex Spark is optimized for speed using Cerebrus inference chips, allowing it to perform tasks like designing a website in just 18 seconds. It's considered one of the most advanced coding models available today.
Why are ads being introduced in ChatGPT?
Ads are being introduced in ChatGPT to open new revenue streams for OpenAI as it moves towards an IPO. The ads are clearly labeled to maintain user trust and aim to democratize access to advanced AI models.