Runway Gen-3 AI Videos Are Kinda Blowing My Mind!

Runway Gen-3 AI Videos Are Kinda Blowing My Mind!

by Matt Wolfe

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Matt Wolfe 👥 1.0M 📅 July 1, 2024 ⏱ 20 min 👁 105K 📄 expert opinion 🧭 2026-08-28
Available in: English (current) Français

Keywords

Runway Gen-3AI videotext-to-videogenerative AIprompting

Summary

In this video, Matt Wolfe provides a hands-on review of Runway Gen-3, a new AI video generation model. He begins by noting that while Gen-3 is not yet publicly available, he has early access through Runway’s Creative Partner Program. He demonstrates the generation process, showing that a 10-second video takes a few minutes to create. He then showcases several examples of what Gen-3 excels at, including abstract concepts, time-lapses, and certain cinematic shots, praising the color palettes and the quality of some outputs. However, he also highlights significant weaknesses, particularly with human hands, which often morph or disappear, and with text rendering, which is inconsistent. He notes that longer text prompts often fail or produce errors. He also mentions that the model refuses to generate content involving celebrities or copyrighted characters. He compares the results to earlier AI video models, noting substantial improvement over the past year. He concludes that while the showcased examples on social media are often cherry-picked, Gen-3 is a significant step forward and will likely be publicly available within a few weeks. He also shares a tip for using ChatGPT or Claude to generate better prompts.

190 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides valuable, hands-on insights into the current capabilities and limitations of Runway Gen-3. The creator’s approach of testing various prompts and showing both successes and failures offers a realistic view of the technology, countering the often cherry-picked examples seen online. The argumentation is solid, based on direct experience, and he acknowledges the subjective nature of his assessments. He also provides practical advice for users, such as using AI to craft better prompts.

Scientific Rigor, Source Quality, Title Accuracy

The video is a personal review, not a scientific study, so the rigor is moderate. The creator does not cite external sources beyond the Runway prompting guide and his own website. The title accurately reflects the content, as the creator expresses genuine excitement about the technology’s potential. The video is well-structured and transparent about the limitations, which adds to its credibility.

150 words

Title / Content Match

The title accurately reflects the content, as the creator demonstrates genuine excitement about the capabilities of Runway Gen-3 while also noting its flaws.

Quality & Reliability

7/10

The video is a hands-on review by a known AI tools commentator, showing real-time generations and discussing limitations. However, it lacks systematic testing and relies on anecdotal evidence, with no external verification of claims.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Concurring Sources

  • Runway Gen-3 Alpha Prompting Guide — Official guide that aligns with the video's tips on prompting.

Contribution & Novelties

The video offers a practical, first-hand evaluation of Runway Gen-3, providing a realistic perspective on its strengths and weaknesses. It contributes to the discourse by demystifying the hype around AI video generation, showing that while the technology is impressive, it still has significant limitations. The creator’s method of testing various prompts and showing the results, including failures, is valuable for potential users.

Pour aller plus loin :

101 words

Radar Profile

The radar profile shows high scores in quantity and quality of information, reflecting the detailed hands-on testing. The technical level is moderate, suitable for a general audience. The reliability is good but not perfect, as the review is subjective and lacks external verification.

Reliability 7/10

💬 Positif. Sur les 30 commentaires analysés, la majorité exprime de l'enthousiasme pour les capacités de Gen-3, tout en reconnaissant ses limites, et certains partagent des astuces ou des demandes de tests plus poussés.