Livepeer Co-Founder Doug Petkanics recently laid out a comprehensive pathway toward a bold new future for Livepeer: one where the network joins the Artificial Intelligence (AI) revolution by providing accessible, low-cost compute to power the expansion of video-specific use cases.

Now tokenholders have a chance to help make that future possible, by voting to fund a special purpose entity (SPE) dedicated to making this vision a reality. Voting will remain open until January 21st.

GPU scarcity and soaring demand are making it harder for startups and smaller companies to develop and run AI models. This raises the risks that AI’s future will be dominated by deep-pocketed big technology companies that charge exorbitant rates to those who want to access computing muscle.

Livepeer is committed to staying true to its core mission of building the world’s open video infrastructure. This commitment to accessibility extends to AI, and to doing all it can to enable the creation of open AI infrastructures and video AI compute as a counterweight to attempts by monopoly players to dominate this emerging market.

AI is expected to have as dramatic an impact on video, as it will on a host of other industries. And with deep expertise in the sector developed over the past six years, Livepeer is in prime position to be at the vanguard of that revolution.

AI has the potential to drastically reduce the time, cost, and expertise required to make top-notch video productions. A single user can now create video clips simply by typing a text or image prompt into a computer.

Generative video won’t replace professional productions, but it can create tremendous cost savings at various stages in the process. It can also make video more accessible and interactive for audiences by enabling features such as frame interpolation and subtitle generation, automatic object detection, masking, and scene-type classification.

We are already seeing the launch of some exciting open source foundational models such as Stable Video Diffusion, ESRGAN, and FAST, among others. By helping these models – and those yet to be introduced – compete with closed proprietary systems, Livepeer can help prevent larger organizations from dominating future innovation.

Should its funding be approved, the video AI compute SPE will take on the important role of validating Livepeer’s cost-effectiveness and reliability as a provider of AI video compute, and showcasing this new capability to the world.

To that end, the SPE will develop the following:

• A fork of the Livepeer nodes dedicated to AI capabilities.

• A subnetwork that node operators can form to perform these tasks with payments cashed in on Livepeer mainnet.

• A frontend app that demonstrates these capabilities to consumers.

• The capture of data that demonstrates the cost-effectiveness of the Livepeer network for performing AI inference at scale and the showcasing of this data to the world.

AI video compute tasks will ultimately become job types on the network, just as transcoding video is right now. Future actions by the SPE will be aimed at bringing the capabilities demonstrated in the initial validation exercise onto the network, and opening the network to a wide range of new job types.

Check out the #ai-video channel in the Livepeer Discord to catch up on the discussion about this initiative and its importance to the future of open AI infrastructures and video AI compute.

We have an exciting road ahead of us. Come play your part by casting your vote today!

Livepeer Co-Founder Doug Petkanics recently laid out a comprehensive pathway toward a bold new future for Livepeer: one where the network joins the Artificial Intelligence (AI) revolution by providing accessible, low-cost compute to power the expansion of video-specific use cases.

Now tokenholders have a chance to help make that future possible, by voting to fund a special purpose entity (SPE) dedicated to making this vision a reality. Voting will remain open until January 21st.

GPU scarcity and soaring demand are making it harder for startups and smaller companies to develop and run AI models. This raises the risks that AI’s future will be dominated by deep-pocketed big technology companies that charge exorbitant rates to those who want to access computing muscle.

Livepeer is committed to staying true to its core mission of building the world’s open video infrastructure. This commitment to accessibility extends to AI, and to doing all it can to enable the creation of open AI infrastructures and video AI compute as a counterweight to attempts by monopoly players to dominate this emerging market.

AI is expected to have as dramatic an impact on video, as it will on a host of other industries. And with deep expertise in the sector developed over the past six years, Livepeer is in prime position to be at the vanguard of that revolution.

AI has the potential to drastically reduce the time, cost, and expertise required to make top-notch video productions. A single user can now create video clips simply by typing a text or image prompt into a computer.

Generative video won’t replace professional productions, but it can create tremendous cost savings at various stages in the process. It can also make video more accessible and interactive for audiences by enabling features such as frame interpolation and subtitle generation, automatic object detection, masking, and scene-type classification.

We are already seeing the launch of some exciting open source foundational models such as Stable Video Diffusion, ESRGAN, and FAST, among others. By helping these models – and those yet to be introduced – compete with closed proprietary systems, Livepeer can help prevent larger organizations from dominating future innovation.

Should its funding be approved, the video AI compute SPE will take on the important role of validating Livepeer’s cost-effectiveness and reliability as a provider of AI video compute, and showcasing this new capability to the world.

To that end, the SPE will develop the following:

• A fork of the Livepeer nodes dedicated to AI capabilities.

• A subnetwork that node operators can form to perform these tasks with payments cashed in on Livepeer mainnet.

• A frontend app that demonstrates these capabilities to consumers.

• The capture of data that demonstrates the cost-effectiveness of the Livepeer network for performing AI inference at scale and the showcasing of this data to the world.

AI video compute tasks will ultimately become job types on the network, just as transcoding video is right now. Future actions by the SPE will be aimed at bringing the capabilities demonstrated in the initial validation exercise onto the network, and opening the network to a wide range of new job types.

Check out the #ai-video channel in the Livepeer Discord to catch up on the discussion about this initiative and its importance to the future of open AI infrastructures and video AI compute.

We have an exciting road ahead of us. Come play your part by casting your vote today!