AMA Highlights — Pastel Network
On Wednesday 13 January, we welcomed Anthony Georgiades from Pastel Network into the Hillrise Capital Telegram Chat for an AMA. Anthony is a cofounder and early investor in Pastel Network.
Built to improve upon some of the core weaknesses of current NFT architecture on networks like Ethereum and IPFS, Pastel Network is a platform catered to digital artists and NFT collectors.
With Masternode componentry and robust security for storage of NFTs, duplicate detection and NSFW filtering, Pastel Network present a fully functional blockchain design created for transparent and permanent digital provenance for the nascent NFT industry.
In this post, we have compiled key questions and answers from the event.
Daniel Dal Bello
Welcome Anthony and thanks for joining us, I’m looking forward to another conversation focused around NFTs and digital art! We’ve been exploring quite a few different startups in this corner of the blockchain world over the past handful of months.
I would like to start as always with some introductions and your journey to where you are now with Pastel Network!
Thanks for joining us today, looking forward to learn about your NFT insights!
Great, hi all!
I’m Anthony, one of the co-founders and early investors of Pastel. I run the project/product management of the platform and much of the day-to-day operations of the business. Jeff is the co-founder and brains/vision by the platform.
In a nutshell, Pastel is all about empowering artists and art collectors to create, collect, and trade provably rare digital artwork in a truly decentralized, secure, and robust way.
When we started this project a few years ago, we concluded that the only way to do this was to make an application specific blockchain (rather than an ETH token) so we could control every aspect. The biggest issue with other methods is that they do not provide storage for the image files themselves, and you can’t put this data directly into the blockchain without it quickly running into scaling issues etc.
Daniel Dal Bello
Purpose-built solutions like this are always interesting for us to look at in-depth.
Exactly, the idea being that It is a purpose-built, self-contained system. Not a general piece of underlying infrastructure that might be able to host thousands of various applications (each with different performance needs) but rather centered around provably rare, digital assets.
Daniel Dal Bello
We have seen over the past several months NFTs and NFT-inspired gaming platforms launching left and right, there has been a lot of buzz around digital art also. However, Pastel isn’t a new project that you’ve built to capitalize on this momentum, it’s been in development for several years now.
What motivated you to start building a platform like Pastel and so relatively early compared to many other successful names in the industry?
I partially answered this earlier, but a key driver in the initial driver and vision had to simply do with the inefficiencies we saw in the traditional market which impeded much of the digital art world from day one.
You’re talking about a global art market of US $50bn+ each year, and one that is largely concentrated around or controlled by a number of high-end artists, galleries, and collectors. There is really no democratization in the market — it’s virtually impossible for new artists to monetize or make a career without initial resources or an ‘in’.
And on top of it, there are tons of intermediaries who take from both sides. Granted, they provide valuable resources. But much of this can be eliminated.
So really, we had the early conviction from day one that these inefficiencies (existing in the market for physical artwork) could be addressed and fixed by a decentralized, trustless system in the digital art world. Which is exactly we set-out to build.
I have a question regarding economic models for artists.
One time sales commissions are the standard in the blockchain/cryptocurrency art and NFT space, are there any other economic models that you’re looking at to implement within your marketplace? For example, a recurring percentage-based artist fee per transfer of an art asset?
That’s a good question and the recurring model is definitely something we’ve explored. In this case, this acts as sort of a “commission artists” model (basically on a monthly basis etc.) to continue supplying art.
The percentage commission is of course the most natural/tightly aligned with the broader platform. We of course are also open to ideas.
Happy to hear that, did you come across any other revenue models that might work well?
We have explored ways for the platform to potentially also act as a ‘promotional gallery’ might to help younger, more emerging artists. But the point of sale (percentage commission), recurring model (recurring fee), and commissioned art model (dollar fee) are the most obvious.
Daniel Dal Bello
If we distill the vision of Pastel into a few words, you have created an open-source decentralized platform and blockchain catered to digital artists and NFT collectors.
How do you position yourselves against the likes of Rarible, SuperRare, or even WAX?
Yes this is exactly right. But keep in mind — it is our system and infrastructure, on a Bitcoin-like blockchain (I can describe the architecture in more depth if you’d like 😄) which is important because we are not beholden to any other platform.
In terms of how we position ourselves to other platforms — they do for sure have certain advantages for both creators and speculators by being interoperable across various across ecosystems.
This can improve liquidity and ‘tradeability’. And then you have specific token standards (e.g. ERC721) which promise provable scarcity.
But if you look at a lot of these alternatives, they either are built upon Ethereum or other blockchains (e.g. Tron). And so the purpose-built application is totally dependent on the native chain.
Imagine an artist starts to take off in popularity or a number of new users quickly join the network and there is suddenly heightened NFT demand. Or even demand for the network by a totally independent, separate application. This traffic would drive transaction fees to a point in which it becomes economically in-viable to purchase a single piece of art.
Additionally, these platforms or even native, so called self-contained platforms like WAX, simply lack the sophistication or network capability to handle the features we have such as deep learning based duplicate detection, distribute storage, or simply scanning for NSFW.
Can you talk to us about your Pastel ID system, authenticity, and how you are able to keep intellectual property secure?
How do you provide assurance that duplicates and theft of artist intellectual property does not take place on the platform?
Also, how do you approach filtering for inappropriate content, and above and beyond NSFW-style filtering, do you have any further ability to censor content?
So the artwork registration flow is essentially as follows. First, an artist simply downloads the Pastel Wallet which generates a wallet address. Then, the artist can create a new Pastel ID, which is a public/private key-pair (specifically, it’s based on the EdDSA protocol using Ed-521 curve 🙂).
This Pastel ID is the ‘core’ for identifying the artist or collector and the underlying piece of art.
Once the Pastel ID is established, they can upload an art file which then goes through the registration process. In this case, it runs through our duplicate detection system, which leverages advances in machine learning and certain statistical techniques. More specifically, it is actually a ‘near duplicate detection’ so things like cropping, scaling, adjusting color/contrast, adding random noise or items etc. will also get picked up as duplicates.
Once the duplicate detection is passed, the NSFW check occurs. We currently use a pre-trained Tensorflow model from Yahoo (OpenNSFW) to assign a score from 0.0 (definitely not NSFW) to 1.0 (definitely NSFW). Once the NSFW check is passed, then 3 randomly selected Masternodes will approve and sign the registration ticket.
The registration ticket is written to the blockchain and includes the artist’s Pastel ID.
Daniel Dal Bello
I always find myself coming back to this question with every good solution and platform that we see.
How will you attract artists and collectors to your platform?
We could draw parallels here to Ethereum and the numerous Eth-killers, which in many cases are objectively better in performance and tech, but have never been able to achieve Ethereum-killing traction.
Well, I wouldn’t necessarily use the concept of Eth-killer in this case because I still think the market is very nascent/young (digital art platforms that is) and no one has really captured a meaningful part of the market.
Sure, a handful of NFT platforms or dApps on ETH have some strong traction, but that is still quite in its infancy and small relative to the market opportunity.
Look — at the end of the day, there is always going to be a chicken and egg problem but I think what is able to really differentiate us from other competitors or platforms is that ours is custom-built, purpose-driven with the underlying user’s (both artist and collector) needs in mind.
For example, the sophistication of the art-registration and Pastel ID is super important for an artist to be able to feel as though their piece of work is truly secure and will not be pirated/copied etc. on the platform.
Likewise, this should give collectors the confidence to attribute strong value to an underlying piece of work and maximize the willingness to pay. The incremental demand for art will drive artists to the platform and increase supply, driving new collectors/traders/liquidity etc so a very strong in-system network effect.
Daniel Dal Bello
Can you talk to us about the Pastel Growth Fund, how you have used those funds to date, and what future expenditure plans are?
Is this how you will fund your ‘Artist Onboarding Strategy’?
The Pastel Growth Fund is really in place to ensure the continuity of the technical development, operations, business development, and marketing of the platform.
For example, our technology roadmap in terms of development include many key milestones such as rolling out a web-based (vs. desktop wallet) application and then support for various file types (videos, audios, GIFs, etc.). We have specific technical implementations we want to tackle like Loaded POW or the use of steganography to encode SPHINCs signatures as QR codes in images.
Simultaneously, we want to allocate funding towards our artist onboarding as you alluded to. This includes commission of campaigns by tackling specific art segments of the market (for example southeast Asian ANIME art communities).
Daniel Dal Bello
So, how are you going to attract quality artists to the platform to kickstart the network effects?
If only I had a magic ball for that, that is always the biggest challenge. Basically, the art market is so fragmented both in terms of the types of art and the geographies/demographics etc — which is a challenge, but also an amazing opportunity.
We have been basically ‘attacking’ the market in terms of these communities, trying to manage each with almost ‘ community members ‘ and running specific campaigns or promotions for each.
For example, a ‘promotional gallery’ approach to ANIME or commission artists for digital contemporary art. And then, Pastel quickly becomes a platform for categorized types of art.
Imagine being able to ‘visit’ galleries on the platform for all of these categorized types.
Daniel Dal Bello
How do you as an entity monetize your platform?
Well, we have full faith and belief in the system and certain of us are digital artists ourselves so first and foremost creating and selling art. In addition, staking PSL for masternodes and mining.
As a last question, we haven’t come across Masternodes nearly as frequently as in years gone by. Can you talk us through the economics of your Masternode design and why you have chosen this path?
I’ll start with why we chose this path.
Think about the complex services we’ve discussed — Duplicate Detection via advanced ML, TensorFlow based NSFW, complex storage, etc. There are off-chain services that require tons of CPU/GPU power. So to deliver these services in a decentralized way, the best approach is Masternodes deploying servers with certain CPU requirements to run the software.
And in this way, anyone who wants to can buy enough of the coin (in this case 1mm PSL) and start their own Masternode and actually run the system — you don’t need permission from anyone, and it means that the system will never stop running because we got bored of it and moved on (or even if we got hit by a bus!) — as long as anyone wants to they can run servers themselves. The block reward is 1,250 PSL/block distributed across Masternode holders.
Hillrise Capital is an independent research and advisory firm exclusively focused on blockchain startups and crypto-assets.
Pastel Network are building an open-source decentralized platform and blockchain catered to digital artists and NFT collectors.