Look At What We Can Do With AI Video Now!

Look At What We Can Do With AI Video Now!

🎙 Matt Wolfe 👥 1.0M 📅 May 10, 2023 ⏱ 27 min 👁 199K 📄 news review 🧭 2026-08-28
Available in: English (current) Français

Keywords

AI videotext-to-videovideo generationRunwayKaiberNeRFD-IDGenmoPlazmaPunkDecoherence

Summary

Matt Wolfe presents a comprehensive overview of AI video tools available as of May 2023. He demonstrates each tool with live examples, starting with D-ID for animating faces from images, LeiaPix for adding 3D depth to stills, and Meta’s Animated Drawings for turning drawings into animations. He then moves to more advanced tools: RunwayML with its Gen-1 and Gen-2 models for style transfer and text-to-video, Modelscope (free but with watermark), and Genmo for text-to-video and image animation. Kaiber is highlighted for its ability to transform videos into artistic styles, as shown with his daughter’s dance video. PlazmaPunk generates music-synced videos, and Decoherence offers a user-friendly interface for Deforum with audio-reactive effects. Finally, he discusses NeRFs (Neural Radiance Fields) using Luma Labs to create 3D scenes from images, and Wonder Dynamics for replacing characters in videos with 3D models. The video concludes with a reminder of his Future Tools website for more AI resources.

153 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides high practical value by showcasing a wide range of AI video tools with real-time demonstrations, making it useful for creators and enthusiasts. The argumentation is solid in that the creator explains the capabilities and limitations of each tool, often showing both successes and failures (e.g., the ‘rubbery’ effect in Animated Drawings). He also gives context on pricing and accessibility (free trials, beta access). The demonstrations are clear and well-paced, and the creator’s enthusiasm is balanced with honest assessments. However, the video is more of a survey than a deep dive, and some tools are only briefly touched upon.

Scientific Rigor, Source Quality, Title Accuracy

The video demonstrates good scientific rigor in that the creator provides links to all tools mentioned in the description, allowing viewers to verify and explore further. He also mentions the research nature of some tools (e.g., Animated Drawings from Meta) and notes the closed beta status of others. The title accurately reflects the content, as it is a broad overview of AI video capabilities. The creator does not cite academic sources but relies on his own testing and demonstrations, which is appropriate for a practical review. The inclusion of a personal anecdote (father-daughter dance) adds authenticity but is not central to the scientific content.

220 words

Title / Content Match

The title accurately reflects the content: a comprehensive showcase of AI video generation and editing tools.

Quality & Reliability

7/10

The video is a practical overview of AI video tools, with live demonstrations and clear explanations. The creator is transparent about limitations (e.g., Gen-2 beta, watermark on Modelscope) and provides links to all tools. However, it lacks deep technical analysis or independent verification of claims, and some tools are shown with subjective opinions.

Chapters

Cited Sources

  • Future Tools — Main website for AI tools directory, mentioned at the end.
  • D-ID — Tool for animating faces from images, demonstrated at 1:11.
  • LeiaPix — Tool for adding 3D depth to images, demonstrated at 2:45.
  • Animated Drawings — Meta's tool for animating drawings, demonstrated at 3:47.
  • RunwayML — Suite of AI video editing tools, including Gen-1 and Gen-2, demonstrated at 5:26.
  • Modelscope Text-to-Video — Free text-to-video tool on Hugging Face, demonstrated at 8:19.
  • Genmo AI — Text-to-video and image animation tool, demonstrated at 8:57.
  • Kaiber AI — Tool for transforming videos into artistic styles, demonstrated at 11:39.
  • PlazmaPunk — Tool for generating music-synced videos, demonstrated at 14:50.
  • Decoherence — User-friendly interface for Deforum, demonstrated at 16:58.
  • Luma AI — Tool for creating NeRFs (3D scenes), demonstrated at 20:23.
  • Wonder Dynamics — Tool for replacing characters in videos with 3D models, demonstrated at 23:44.
  • Future Tools Newsletter — Weekly newsletter for AI tools, mentioned in description.
  • Matt Wolfe's Blog — Personal blog of the creator.
  • Mubert — Music generation tool used for outro music and in Decoherence demo.

Concurring Sources

  • RunwayML — Official website of Runway, confirming the existence of Gen-1 and Gen-2 as described.
  • Meta AI Animated Drawings — Official demo page for Meta's Animated Drawings, confirming the tool's availability.

Contribution & Novelties

The video provides a timely and comprehensive survey of AI video tools in mid-2023, highlighting the rapid progress in text-to-video and video editing. Its original contribution lies in the practical demonstrations and comparisons, offering viewers a hands-on sense of each tool’s capabilities and limitations. The inclusion of NeRFs and audio-reactive effects adds depth beyond simple text-to-video.

Pour aller plus loin :

  • Neural Radiance Fields (NeRF) — Foundational concept for 3D scene reconstruction, central to Luma AI’s tool.
  • Stable Diffusion — Underlying model for many image generation tools, relevant to the style transfer techniques shown.
  • Deforum — Open-source tool for AI animation, which Decoherence simplifies.
  • Runway Research — Official research page for Gen-1 and Gen-2, providing technical details.
  • Hugging Face — Platform hosting Modelscope and other AI models, relevant to the free text-to-video tool.

133 words

Radar Profile

The radar profile shows a balanced performance across all dimensions, with slightly higher scores in quantity of information and technical level, reflecting the video's comprehensive coverage and practical demonstrations. The lower score in quality of information suggests that while the content is informative, it lacks deep technical analysis or independent verification.

Reliability 7/10

💬 Très positif. Sur les 30 commentaires analysés, la grande majorité exprime une forte appréciation pour la qualité du contenu et la valeur éducative, avec des remerciements pour le travail de recherche et des demandes de continuité.