TL;DR Summary of the AI Litepaper
The AI Litepaper explores the potential to deploy Large Language Models (LLMs) on POKT Network
Another edition of Infracon is in the books! After our first Infracon in the Dominican Republic, Infracon Day Bogota was a chance for some of the top middleware protocols to shine, and to share many of the trends and innovations that we’re seeing in this space.We were thrilled to bring so many thought leaders together to share their insights (we had nearly 20 different sessions!), and to be with our amazing community in-person yet again.Let’s check out some of the highlights and takeaways from a jam-packed Infracon Day. If you're looking to check out the talks in their entirety, head over to the Agenda page and find them all there!
Pocket Network CEO Michael O’Rourke kicked us off by stating an important point: Web3 middleware is underserved.As other speakers alluded to throughout the day, there are still incredible opportunities in areas like indexing, RPC infrastructure, privacy, compute, storage, and AI. These broader areas can have game-changing impacts for more specific use cases: identity, social dApps, games, oracles, DAOs, open finance, decentralized science, and climate data were just some of the use cases that were flagged to keep an eye on. One speaker in particular cited the potential for the first breakout Web3 social media app to act as a major inflection point for the rest of the space, while another referred to decentralized microservices as “the frontier.”
“Solutions are generalized, but use cases are specific.”
Even though there are still barriers to open innovation (including a “separation of trust and innovation,” as described by one of the speakers), Web3 infrastructure is offering new guarantees of ownership and data portability, guarantees of data integrity and verifiability, and permissionless, open access in a multi-chain environment - all of which will power the sectors and use cases above.
To best tap into some of these opportunities, it will be crucial to build diverse teams in Web3, and on-ramp underrepresented communities. Brandon Middleton, with University of Illinois Urbana Champaign, had a great session on this topic, emphasizing the importance of developing skills to build diverse pipelines into Web3.We can get there by helping others to build creative confidence, establishing growth mindsets, and making Web3 more accessible for all. Importantly, this needs to be done while keeping cultural contexts. Towards these efforts, there have already been some exciting case studies at the likes of University of Illinois, Stanford, and UC Berkeley (among others).
Jon Kol, Co-founder at Hyperlane, had a great quote during Infracon Day:
“We are working towards the Interchain Singularity.”
Jon laid out several different eras of interoperability that are taking place simultaneously, and overlapping with one another, including the App Centric Interoperability that we are experiencing now. As we move through this era, developments like interchain accounts (remote identity proxies that allow devs to control addresses on remote chains without deploying contracts on those chains) and interchain queries (viewing calls across blockchains) will play a role, and we will see a continued shift from permissioned to permissionless.
“Permissionless deployment is the holy grail of interoperability.”
While there is still a need for low latency bridging, we are seeing meaningful progress towards trustless bridges, including work being done to shift from simple multi-sig to a light client, make costs reasonable, and provide better security guarantees - particularly with some of the efforts that our speakers from Gnosis Chain shared.
Moving towards greater interoperability makes one thing clear in particular - data is king.Several speakers throughout the day brought up some of the key pain points with data indexing, including the sheer amount of time (sometimes weeks) to query and fetch large amounts of data.Thankfully, we’re seeing innovative data solutions continue to emerge, including different ways to incentivize data providers and validators for high quality data feeds, better ways to manage access and update rights, better dispute resolution mechanisms for data integrity, and optimized querying.These innovations will help us move further towards valued and trustless data, and the idea of data as a “public good.”We even have projects like DB DAO, who is turning databases into DAOs, including making each row an NFT that can be viewed on OpenSea. We will likely continue to see new approaches to monetizing databases through ads, subscriptions, or data set sales, and even data tokens that are backed by IP. Simultaneously, DB DAO is tackling “data pollution” problems with their Curator-focused approach, including making curation a multi-sig. As Michael Fischer described it, they’re helping to stand up “laboratories of data democracy.”Evolutions across the blockchain API technology stack, in areas like compute and storage, blockchain protocols, blockchain nodes, middleware, service gateways, and applications, will continue to solidify data’s position on the throne.
While added layers are progressively decentralizing Web3, there remains much work to be done on this front. Oligopolists are still extracting fat margins and censoring as they wish, and supply chain difficulties are also stemming from centralization. Some of the profound failures in centralized systems are leading to a real, organic demand for decentralization.Decentralization at its core is a transfer of authority and responsibility, but is a complex and multi-faceted concept to tackle. In Web3 specifically, there are four especially important “dimensions” of decentralization: governance, technology, fiscal, and market.Among some of the specific centralized issues to address, some of the Infracon Day speakers included:
As we think about these issues, we may need to review the metrics we use to gauge decentralization. Our speakers from Gnosis Chain, for example, presented the idea of the “Correlated Validator Sets” metric as a more accurate representation of decentralization than the Nakamoto Coefficient.
Unsurprisingly, the work being done towards increased scalability in Web3 was another key focus of Infracon Day.The goal across Web3 is to bring higher levels of security, speed, and scalability to the table, without restrictively high costs. Towards this end, rollups will help scale computational needs and create “abandoned blockspace” that is secure and scalable - but even in a “rollup era” there are still problems with data bandwidth. Finalization times are often too slow, and faster settlements are needed. Infrastructure often doesn’t scale up OR down effectively, and the space as a whole needs to find better ways to scale to millions of users.Gnosis Chain in particular shared some of the work they are doing in pushing the limits of EVM blockchain computational capacity, including growing block size limit, reducing block times, and solving the key problem of state dependency.
We also had our very own Daniel Olshansky moderating a panel on how protocols can grow their demand side. A number of tips were offered up on this topic, including:
Overall, protocols should determine what success looks like, and work backwards to achieve that success. It’s absolutely critical to be mission-oriented, and to build an ecosystem of people that believe in that mission and want to be a part of it. On top of that, it helps to have a project that is uniquely enabled by Web3, and to be able to convey that effectively.The panel touched on a number of other interesting topics, including different funding strategies, whether government buy-in is necessary, and when to use a token.
One aspect of Web3 that needs to remain a point of focus is UX.As Doug Petkanics of Livepeer shared, we are missing a “guarantee of experience” when it comes to UX across much of Web3. This is something that must be overcome to contend with Web2 experiences that users are accustomed to.Specifically, there is still not a viable content delivery platform focused on video in the Web3 space. To address this, we need to work towards more standard interfaces that can achieve that “guarantee of experience,” as well as better infrastructure portability and an open marketplace encouraging healthy competition.
Another recurring theme throughout Infracon Day was the importance of proper incentives in Web3.Even from the perspective of a creator’s economy, we need to continue to find ways to incentivize good content, and should be considering different ways that viewers can share in revenue as well (similar to Brave’s approach).Michael Fischer of DB DAO spoke on this topic from a data perspective in particular, describing the ways in which the revenue that a database generates can be shared. In this way, we can better align the contribution of useful data with real economic incentives.
Finally, there were some interesting discussions at Infracon Day around the evolution of keys and security practices in Web3.Lit Protocol in particular is doing some fascinating work around turning keys into an application development platform, as they’ve integrated several hundred apps already. They are also focused on using decentralized cloud signing (“programmatic signing”) to automate writes to state machines, as well as automations for biometric authorizations for wallets.Remember, if you want to check out any of the sessions in full, head over to our Agenda page where we've got them all posted!We want to give a huge “THANK YOU” again to everyone that joined us in Bogota and via our livestream - this event wouldn’t have been the success it was without sharing it with you!