Irreverent Labs' Text-to-3D AI Video Creator Tools

Creative media, especially in video form, is seemingly never dying down as a form of entertainment in our modern age. With the addition of artificial intelligence continuously booming and being further developed, players in the tech industry creating AI-powered video creation platforms and tools keep expanding.

Many of these players build their technology with the good intention of making video creation accessible to everyone, even to those with no prior experience or knowledge in creating videos. However, one of the biggest sources of concern often encountered in the field of AI is the misuse of existing works being used to train models without compensation or acknowledgment to original creators.

Founded by Rahul Sood and David Raskino, Irreverent Labs aims to democratize the birth of entertainment in the form of 3D videos with its AI-powered tools while still applying safeguards to protect intellectual property rights.

Pursuit of Democratized Video

Prior to its pursuit of AI-powered video creation tools, Irreverent Labs, founded in 2021, focused on creating an NFT-based robot cockfighting video game titled MechaFightClub. After a shift in the US’s regulations toward cryptocurrency, Irreverent Labs shut production down.

While developing the game and, in the process, learning to animate fight scenes and build characters, Irreverent Labs’ research team discovered a new way of utilizing the AI-powered technology it had developed, and this was the beginning of Irreverent Labs’ video creation foundational model’s fruition.

Irreverent Labs developed its foundational model after raising $45M in a funding round led by a16z’s Andreessen Horowitz last year, using the funds to purchase data-storing GPUs. Filled with ambition, the founder duo continued building its tech in the hopes of giving millions of users worldwide the ability to create videos with simple natural language prompts.

The dream then came closer to reality when Irreverent Labs closed a financing round from Samsung Next as a way to explore a distribution channel that the founder duo believed wouldn’t have been possible if the deal hadn’t been done. With plans of not only creating a video creation platform but an API as well, Irrevent Labs plans to reach out to Samsung Next’s portfolio companies once it has completed production.

Magic in Video and Secured Property

In its tech creation process, Irreverent Labs utilizes a hybrid approach that mixes video analysis, synthesis, and entropy models that the startup believes could lead to high-quality, more realistic video outputs for users.

Upon research, Irreverent Labs’ teams found what’s called a “temporal consistency” as one of the obstacles found in AI-generated video creation. What this means is that pixels occasionally have trouble figuring out where they need to be, which can potentially result in poor-quality video outputs. The startup’s hybrid approach is used to avoid exactly that and so that videos made using its Irreverent Labs’ platform do not seem AI-generated.

Irreverent Labs has also grown aware of the type of trouble AI brings to the creative industry. The startup is in pursuit of a solution to protecting intellectual property rights while still continuing its AI tool development, as it wishes not to stray away from its goal of democratizing video creation.

With an alarmingly large amount of misused AI tech for video creation, such as replicating one’s likeness in the form of deep fakes without consent or models being trained on studios’ original works without compensation, Irreverent Labs is in the process of creating guardrails that final video outputs to be finetuned according to customer needs.

Irreverent Labs’ AI model still has to be trained on publicly available data from the internet, rendering the chance of training on existing works extremely possible. To avoid intellectual property rights violations, the startup plans to make several policies, one of them being letting studios whose character IP’s likeness is used on a video created using Irreverent Labs’ tools monetize the video outputs.

Upon completion, Irreverent Labs plans to reach a distribution deal with Samsung that would allow the startup’s foundation model to be integrated into Samsung devices across the world, used to turn users’ photos into short-form video.

Samsung Next investor Joan Kim is optimistic about Irreverent Labs’ potential impact as their collaboration goes on, with the chance of AI-powered being accessible on phones and televisions, easily within our reach.

Meme & AI-Generated Picture

Job Posting

• Epam Systems - Azure DevOps Engineer - San Francisco, CA+ (Remote/Hybrid)

• Grammarly - Lead Product Designer, Core Product - San Francisco, CA+ (Remote/Hybrid)

• RapDev - ServiceNow Solutions Engineer, Sales Engineer - Boston, MA+ (Remote/Hybrid)

• Self Financial - Senior Analyst, Product Analytics - Austin, TX+ (Remote/Hybrid)

Promote your product/service to Digger Insights’ Community

Advertise with Digger Insights. Digger Insights’ Miners are professionals and business owners with diverse Industry backgrounds who are looking for interesting and helpful tools, products, services, jobs, events, apps, and books. Email us [email protected]

Your feedback would be greatly appreciated. Send it to [email protected]