Community Call - 11/5/18 - 12:00pm ET, 6:00pm CET, 9:00AM PT

[dob]

Participants and location

The community call is open to all.

Participant Link: meet.google.com/wqt-putu-xwo

Live stream link to watch and follow along if not planning to actively participate: https://livepeer.tv

Agenda

Devcon4 Recap .

Streamflow Update Livepeer has begun work on the engineering for its next round of protocol updates, called Streamflow. Hear about the first steps in building it out.

Open Q&A and Discussion

Please add any questions or topics that you'd like to discuss in advance to this issue.

Roles

Moderator: @dob
Note taker: @ericxtang
Streamer: @angyangie

Anybody should feel free to add agenda items or leave questions they'd like to discuss here in the issue. I'll keep the agenda up to date.

[dob]

Here are the notes from today's call:

Devcon 4 Recap

Takeaways from Devcon as it relates to Livepeer were:

• 50%+ awareness of Livepeer, relative to ~10% awareness at Devcon 3.
• Potential collaboration opportunities with other web3 infrastructure projects and teams who are looking to get involved in the network.
• GPU transcoder mining thesis was partially validated via feedback. GPU operators are looking for additional revenue streams and are willing to work with Livepeer on one offs to test the viability of transcoding on their clusters. Need to investigate/determine their bandwidth costs.
• The generalized mining trend is picking up steam with entrepreneurial and investment minded participants, and Livepeer is the case study on Ethereum of a network with active participation.
• Streamflow review and feedback is best received when separating out PM and payments architecture discussions from staking economics discussions. Good feedback in each area, but trying to get someone to absorb all of Streamflow at once is too much.
• Ethereum scaling solutions are emerging, but many silver bullet solutions still feel very far out, and I'm confident in our decision to take matters into our own hands with offchain scaling rather than waiting for other research and projects to deliver.

Streaming on Livepeer

There's an abundance of inbound requests to stream events on Livepeer, except Livepeer isn't an event production operation, and the core development team can't fulfill these requests. Instead:

• Route high end enterprise style requests to partners who know how to use Livepeer's network alongside redundant backups. They additionally have great customer support and operations to make sure the full service of the events go great.
• In the short term, we either have to route grassroots requests towards the DIY option, or unfortunately can't spend time on core team support.
• In the medium term, there's a tremendous opportunity for a community-driven, Open Streaming Collective, to emerge and support all of these grass roots requests. This would entail the community operating under a separate brand, and being automomously responsible for the end user experience and applications - though they could certainly build on Livepeer and be a flagship DApp/partner.

In short, when events have an expectation of very high quality produced content, recording, customer service, etc - it's not right that the world thinks "the event is streamed by Livepeer" - because Livepeer is not an event production company that can deliver on that full experience. When things go wrong, and they will, it reflects poorly on the project due to things outside of the project's control. It's better if the world thinks "the event is streaming using the Open Streaming Collective", which can facilitate that. And the OSC can be a built on Livepeer, in addition to other services/protocols that facilitate a great event streaming experience.

Q&A

How does Livepeer know that it's staking economics are working?

• We should be more disciplined about the weekly economics updates, which report on the observations of what's happening week-to-week.
• Current staking is ahead of projections and outpaces the rate of vesting+inflation, so people are actually responding to incentives.
• For Streamflow, it would be awesome to use agent-based simulation to predict what will happen based upon various parameters. There are research areas to explore here, but we'll see if it's practical.

How does Livepeer prevent stake concentration on the top nodes?

• In Streamflow, given equal pricing and fee share, stake should flow towards the nodes who are doing the most work and collecting the most fees, that are currently understaked.
• This will result in an equilibrium where stake is widespread across all the various services that are fulfilling open demand, in rough proportion to the fees that are being generated and shared amongst those nodes.
• Currently you're looking at an early stage bootstrapped network where most of the incentive is inflation, and in this case it is hard to prevent the stake concentration.