Filecoin FVM Space Warp Hackathon, 25 Jan 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: FVM Space Warp Summit 🛸 Video Use Cases on FVM

Description

Sarah Thiam of Protocol Labs joins Yondon Fu and Akshit Gupta in discussing video use cases on FVM.

Follow us! https://linktr.ee/filecoinio
Subscribe to our newsletter! https://bit.ly/3Oy8J9j

#fvm #filecoin #fevm

A

A

And we are up into our last but of course, certainly not least top, uh so we're gonna welcome back uh Sarah to the stage and Sarah's going to be the moderator for uh a chat about video use cases on the fvm, so video, of course, being an incredibly data. Intensive application.

A

uh We're joined by two awesome Builders uh in the space, so we're joined by action from cuddle01 and yonden from Alive live peers, so I'll also invite them up to the stage and to unmute and share their screen or share their video as well and I. Will let Sarah take it from here.

B

Thank you, Jacob all right, cool, hey nice to meet everyone again, um we're back we're back with more um teams that are building uh around with and in the fall Point ecosystem. um Today, we're joined by live here in huddle. One um really exciting teams that you know we've been talking a lot with. um We've also done like some amazing amazing events with um the teams.

B

Like you know, on Bungalow, or with really huge for in the blockchain week that was super exciting, had all one did a meeting for Bungalow, um so yeah I'm really happy to have both of you come on today. um Today we have yondon and akshit from knife pure and had one so I'll. Let you do a quick introduction uh for those of you who have not met me in the previous talk, I'm Sarah from the femt underdeveloper advocate, so yeah I will let down London actually do a quick introduction.

C

Sure I guess I can kick things off. So thanks for the introduction and thanks for having me, my name is Yonder and I'm. One of the co-founders of Life peer and these days, I lead a lot of our engineering and research efforts and yeah happy to be here and to talk about video and fvm and like.

D

So hi guys so I'm akshit I'm a full stack developer at hurdle. Zero. One and I lead the front-end team there and I'm. Also one of the founding Developers. Is it so? The teamwork revolves around the SDK so which powers that fall yeah.

B

Yeah thanks thanks for that great introduction.

B

um We would love to hear much more about what I think for those who are new um to what not even had a one um offer and what they're building and we would love to hear more about what you're building these days, where you're really excited about both of you are um come from teams and projects that are very video Centric. It is you know, video is huge um in general, you know web 2 web 3, regardless um it's really really huge right.

B

Now, it's like being a torment and look popular, um but decentralized videos are going to be a really huge thing right. uh What we want to find out the day from you is: what are you building where you're excited about and yeah? Why why a decentralized version of video um so yeah I'll London? Maybe you want to start off with um Wildlife peers up to the GT and what life care is about for those that are new.

C

Sure so, taking a step back, I think at a high level the way that I've described the work that we're doing at life here is Building, open and decentralized video infrastructure, and we decided to focus specifically on video, because uh the reality is that video being super data intensive and compute heavy is one of the hardest mediums of content to scale on the internet today, and it's pretty cost per hit at it, and a crucial piece of this infrastructure is video processing, uh one flavor of which we've focused on.

C

Is video transcoding the process of taking a single video and basically being able to translate it into all these different versions, so that you can effectively play it on any device anywhere around the world and back compute is pretty heavy, and we've created this Global decentralized Network that offers this service and an Open Marketplace, where you can bring your hardware and bring your bandwidth in order to push down this cost of service and provide this service to the network uh in an open and decentralized manner uh and the way that we work with developers historically.

C

Is we have this selector Studio product, that's built on top of the network that offers a Gateway API, an SDK, as well as various video player tools, to make it easy to integrate, live streaming or video on-demand experiences into your app, so I'll pause there and happy to expand further later on.

B

For sure um akshit, how about you and hello one.

D

Yeah absolutely so Hydro zero one is a communication toolkit for web3 and our mission is to cater to 100 million members who are hard to zero one by 2026, and uh we have three pillars which are uh motion is about, and uh first is the Hydro zero one web platform, which is a sufficiently decentralized real-time communication platform for nfts and DOW communities and web 3 native users. So uh if you have a Dao or in an SD Community, you can create your calls user nfts as PSPs.

D

So you feel like feel right at home in the web c world, then we have our sdks. So we have Plug and Play toolkits for Builders to you know power the tabs with Audio Video Communications in their product with a few lines of code. So even our web platform is powered by our SDK infrastructure. So that's a proof of how powerful our web sdgs are and using our sdks a lot of hackers even built um beautiful products uh in some of our hackathons.

D

We recently had one December holiday hackathon, in which we saw beautiful products and even one with light beer called incur, uh which was uh pretty awesome. Then uh the third winner at Hydro, zero one, is 3D RTC protocol, which is called Divergent, and it is a peer-to-peer protocol uh which in which you can leverage uh decentralized communication uh using any RTC client. So, uh even if it is a drtc app like if it is an app which is uh backed by webrtc, you can directly directly.

D

You can talk to our trtc protocol and leverage the infrastructure. So that is a summary about Hado, zero one and what it is bringing to the web 3 world.

B

Paula thanks so much for sharing that I think we, so our team, as well, has played around a little bit with both lecture and puddle 01.. um We I think one of our developer Advocates met like he had been publishing his video.

B

If you live here, and he really wanted to explore more of that, um how we can get and knocking looking at it as well like how can we get all the great tutorials that we're doing and kind of tested our life here and being an integration through fvm um in the future and bottled Rwanda? We also want to convert like I'll check in then I'll call if we're good at that experimenting to try using the platform as well, um and it well I think we thought like a live um update.

B

That was another Phil Bangalore about how they're like at that control now for conferences, Right Live conferences, which is pretty cool because then you can you like learn green will be just you can't actually see who's behind them. um So I think it makes for a very interesting um new feature through all these live conferences and live meetings that we're doing um I. Think one of the next question that I have for both of you is um what are some ways in which you're exploring working together with the Falcon ecosystem and and in particular fbm.

B

If there are any plan.

C

Sure so, I think the most relevant on the live peer side at the moment is we're pretty excited about supporting and advancing video streaming on the internet, uh with decentralized protocols and decentralized networks, and lately we've uh been really focused on trying to make video streaming work really well with ipf at Nissan drive, storage protocols like ipfs and filecoin and I think uh what some people might be aware of, but not many that are outside of this particular domain.

C

Is that there's just so many different parts of the video pipeline that are stitched together in order to make video streaming work at scale? It's not just storage, it's not just compute. It's not just fast retrieval and content delivery. It's really all of those things under the sky integrated in a tightly uh slightly conjoined manner, so that things can work really well.

C

So we've spent a lot of time over the past couple years, really making sure that we can push the cost of processing down and we've also built end-to-end services that leverage that network, but we're also really interested in basically taking the capabilities that we've built and supporting video streaming within the ipfs and Falcon ecosystem.

C

An example of this is uh in the fall we launched this ipfs playback feature where you can take any uh video CID that is located or stored in ipfs on the public network or on filecoin, and to take that CID transcode it into all these different outputs and Renditions and stream. It back for anyone using that same CID, so basically giving developers that are used to working with cids a similar interface but enhanced capabilities. And now you can stream uh at a on a more scalable Banner.

C

So we're excited to continue working with the ipfs filecoin teams. The folks working on treble in order to make this say a first class feature of of the networks that we're a part of as a dates to the fvm I think.

C

Maybe we'll have time to talk specifically about a more specific applications and use cases, but generally I'm excited about the prospect of these programmable storage contracts that enable these collectives to come together in order to provision, storage and compute resources, because I think that really changes the Paradigm of computing and cloud computing that we're used to where you'll have a single individual rent resources from someone that has the resources required in order to scale your app.

C

But now we can do things like a collective or a community comes together and they can jointly using smart contracts, rent resources, provision resources together in order to create interesting products that create value for that community. So that's like my super generic answer, and maybe we can talk about use cases or apps separately.

B

Yeah yeah for sure um I think it's a lot of exploration right now. Fpm I mean at the old Lenola and all of you, both were know like the product is being built as we go along. So there's a lot of explosions to do that. um Cracking retrieval data retrieval that is huge um and really really important task that we're looking at and the fact that life period managed to crack it for a huge, like video being such a huge like um the NFL as well.

B

I think that super interesting um well, yeah I, would like to check out more on the ipf side of things and how how that, like the new feature that you mentioned, um is actually working right now, yeah, um how about you actually like how? How are the green paddles number one.

D

Yeah so um title zero, one from the start. You know leverages ipfs and file, one for storage, so make all the meeting recordings. And you know the live stream recordings which are powered by Nike are stored on the five coin Network. So uh once we have, you know this programmability that uh you want.

D

One just talked about uh we'll be able to compute over this data, and you know create much more powerful maps and we can, you know, create marketplaces for videos uh which which could then in the users, could even uh curate their content over there and do multiple things and you uh another use case that we are even researching into was storing our metadata using uh fvm actors. So uh we even uh I'm actually in the file coin, foreign.

D

Program where I'm also reading about those getting it up to date, so that uh we can implement the FM things in Hardware zero, one as well. Yeah.

B

Yeah for those of you who missed the earlier panel um The Foundry program, is the fcm early uh sem Foundry early Builders program, so we have like teams that come on. um They maybe have an idea. Maybe they have to exploring and they built together about for the period of the milk Moon.

B

um So that will be going to mean that launch they actually have both projects to go out and that's what huddle um01 had been participating in um yeah and, if, like Pierre, wants to join very happy to think you want to um yeah I called I mean thanks for sharing those today, um you know it's something that um yeah.

B

Well, we definitely a lot of opportunity to build that on, um given that programmability just will bend up by really anything that you can imagine um in it, then uh I guess the next question that we we do want to know and I I know a few people from the community, um and this is the general thing we went to and what do you right, especially for video being so robust uh in web 2 and a lot of people consuming off what two platforms right now? uh What do you think from your perspective?

B

What is the reason that moving to a decentralized video platform, um what what makes it more interesting or better or more efficient or more performant um than what two options that exist right now.

C

Yeah so I guess um always worth uh kind of talking about the distinction between uh say, like the application platform and the underlying infrastructure. uh I working on the infrastructure side can speak a little bit more to the infrastructure, but there's uh certainly reasons to uh favor uh web3 application platforms as well, but on the infrastructure side. From my point of view, I think there are two main things in my mind that um web3 infrastructure and decentralized infrastructure represents.

C

The first is a really big opportunity to commoditize uh existing bandwidth and compute and storage resources in order to drastically push down the cost of this infrastructure with Community powered networks and Community powered marketplaces and video infrastructure is just such an expensive offering in a traditional cloud computing context where, for that reason, it's actually really hard to bootstrap new new creative applications, because in order to even get to a reasonable scale, you you're dealing with all these fixed and variable costs that are massive uh right off the get-go and, as a result, it really leaves you with very few business models that you can explore uh and in our point of view at live here.

C

We think that there's all a lot of constrained uh constraints placed on application, Level Innovation, when uh the only avenues that you have to explore are dictated by what your super high cost basis is in provisioning infrastructure, so I think from a cost point of view with the decentralized network. With these token incentivized networks, we have an opportunity to really push down costs and that just enables you to explore new application ideas that you currently can't uh the other bit. That I think is fairly important.

C

Is this notion of having a credibly neutral playing field for you to build on so with a lot of the infrastructure, services and SAS companies the build on top of.

D

C

While oftentimes, it might make sense for you to start off there uh really there's a lot of variability in terms of things like pricing uh cases, where your infrastructure provider ultimately ends up competing against you by releasing a competitive offering uh because, and they can obviously offer that at a much lower cost basis than you can, because they own those resources. So it ends up with this. Dynamic of you really want to be building on an infrastructure layer where it's not a credibly, neutral playing field, but with decentralized networks and decentralized infrastructure.

C

We do have an opportunity to lay out that infrastructure in a way where it is a Level, Playing Field and you, as a builder, can start building today with the confidence that this will continue to exist and work in the way that you think it'll work in the future, which we think is also quite important for experimentation and permissionless building.

B

Yeah I think I also I mean as a kind of The Benchmark for those that are listening in um I. Think I read somewhere that live here, brings it down to about 50 times the cost ability that, with with traditional or like web 2 existing option for video um yeah for video streaming right.

B

C

You yeah yeah, yeah and I. Think that, and a lot of that is to the first point of uh really leveraging the Dynamics we can create with these networks and tokenize incentives and Community powered networks where uh these resources exist today, let's try to aggregate them and then put them to good use for the people that want to actually build cool things with it.

B

Yeah um and- and you know actually for yourself like how did huddle zero one spot this, um this Gap in in being able to provide decentralized to Doom conferences in a way.

D

Yeah so uh like this actually started in the hackathon itself, so our Founders are you shenzo Schmidt. We were in the hackathon and they saw everyone building their own web 3, but they saw all the meetings being done on Zoom. So that's where the idea for Octave of you know having a decentralized solution for video conferencing and what really uh differs from the web to equivalent of those are like in Google meet and zoom like uh we are doing a meeting right now, so you it doesn't really cater to the website community so like.

D

If you have an NXT or a Dao Community, you would want to see like uh your a part of a group right. So you you can't token get your calls using those tokens. So that's what I will solves. Then it also gives you like crypto Primitives, such as pf3s uh nft phps, which also bring brings in like a better alternative uh for the Webster crowd and uh other than that, like we're. Also leveraging file coins, so filecoin is way cheaper than using other centralized Solutions such as Amazon S3 and like once the retrieval is solved.

D

Then it will be much more uh much more easier for people to retrieve their data and uh use it on there. So these are some of the things that uh differs us from the web. 2 equivalence and other than that like a web, RTC also comes with peer-to-peer calling, so our drtc protocol is also uh peer-to-peer, so it comes with two things. First, it's your data is not being shared to any Central servers. It doesn't go to any Central server and then the other thing is that the latency is also very less.

D

So you don't have to go around central servers and uh to the peer thing, so the latency is very less. So that's how the web 3. It differs from the web to experience.

B

Yeah um and your related question on this and our curiosity, um what what are your thoughts around like being able to maintain the network of different providers, because I think for the dental life Network, especially when it's like with storage and you need? um You need a lot of people. um They are participating in network to be able to deliver the fast streaming time that you need. Is that a challenge that you get and how hard I mean? Briefly just out of curiosity, how do your team deal with that.

D

Yeah, uh it's it's actually a little difficult to do with that, because um right now like we, we have Central Service uh on our main platform, although, uh like we are on a road to Progressive decentralization where these, like the providers, the media server providers will be incentivized using tokens or other markets or other video markets, and uh that's how they'll be able to power that communication. It's a long road ahead.

B

um And I didn't do close off. One of the I mean thank you so much for sharing like more insight into the the back ends and the thoughts and the considerations behind building the decentralized video platforms, um if not anything task at all right and the fact that you are able to watch videos like the way that we are used to it um and it being fully decentralized. It's just surely a meeting. So thank you so much for um building the project that you're building um I, guess to close off.

B

One of the things that we again want to know is um from for both of you. What are some depths or video adapts that you would be really looking forward to seeing from hacker.

C

uh Yeah I I guess I can uh wrap two off first. um So when I thought about this, uh with this question, I have uh maybe like simpler example and uh complex like that would be awesome example. uh But the simpler example would just be uh I think, with the fbm storage contracts.

C

I think it'd be really cool to start, seeing uh things like uh really actual permanent Perpetual storage of certain videos that you want to preserve forever for archival reasons, and it's just sort of like a set it and forget it experience, and you know that this video will always be accessible and it might be a really important artifact from a knowledge standpoint, uh ranging from that to uh you, having like a massive Archive of videos, that you want to preserve and being able to uh program like a large volume discount on that, uh because you know that it's going to be like many petabytes of archival video data that is really important to preserve, from kind of like a human history standpoint or something along the line.

C

So I think those are uh interesting like it's just taking videos and medium, but it's just generally exploring what you can do with video preservation with the fdm contracts and the thing that is probably more complex, but I think it would be really exciting. Is with these programmable storage contracts uh and the research that folks are doing on data Dows, creating these templates so that communities and collectives can spin up a data down and that data now has all the baked in functionality so that it can bundle together.

C

The payments it needs for storage bundle, the payments it needs for compute for retrieval, and then all of those are handled by the Dow itself on a regular basis, so that it Provisions resources for anyone that is a dow member. And if you're, a member of that Dow, you get to use those resources because you're a member of that Collective, and then that gives you the opportunity to explore these modernization models where maybe the collective comes together.

C

Provisions these Resources by bootstrapping some funding and then creates a product of value and then a portion of that value goes back into the Dow's treasury. In order to continue provisioning resources for the Dow going forward, and in this way you can actually start exploring these self-sustaining applications that can pay for resources on their own and then it abstracts a way that need to say, like I need this for storage.

C

I need this for compute I need this for retrieval, and you have these templates that can kind of package things up so that people can just deploy these and experiment uh and I. Think that would be really awesome.

B

Okay, no, it's understanding right now, if the hacker that was the thoroughly a project that you just described so take note and you can build it as well. So thank you for giving the answer. um Okay! Actually, what about you? What would you like to see being built.

D

Yeah so I'll give a reference of a book that I was reading the other day, so it's called the great good place by the uh Oldenburg and he talks about uh unlocking the third place. And you know the third place is a place where people come together and uh you know they come voluntarily and they come technology and anticipate uh together there. So uh this, what this does is that it spawns video bombers.

D

So if, uh like you see these days, uh there are platforms like Tick, Tock Instagram, we see videos all over the Internet, so we do is a really powerful thing and um like I'm, really excited about uh like platforms like video marketplaces, building where creators can you know curate their own videos, exchange them in a decentralized way and uh they can charge it charge for it and there there will be a competition between each of them, so that will also hike down the prices.

D

So uh those are some great use cases that uh of uh programmability. That uh fcn provides and I'm really excited to see things around this, uh where social, adapts and PDO Commerce kinds of things are being built on fcm.

B

Yeah I can totally see like a Marketplace for videos being traded around.

B

um That would be kind of interesting to see how that would look like um and then just throwing in my idea, because I was thinking about it as you both were talking, but I would love to see like uh like a step or like a data dial the desktop titling for videos, um whether it's like detention at Computing, because I know that that's not an easy one to solve, um because usually, if you're doing publicly for live video streaming, you would have like a live translator or you have a really strong AI model, um we'll love to see how that can be decentralized, that they can make a huge difference.

B

The compliment um all the great attention live, video platforms out there today, um yeah the quick little brainstorm, um feel free to take our ideas and build it out, we'll let it be only project, you know being formed out all the discussions that we had um and that definitely continue to be on the difficult um and on the fall coins back Channel um yeah. Thank you so much to your chip for coming on the call today from your different time zone.

B

um That was a really great discussion on what video has to offer in the web. 3 space um and yeah guys, please check out, live here and huddle01. cool. Thank you so much well,.

C

Good. Thank you.

B