Today we announced that we will focus on making NEAR the home for User-Owned AI. We need AI that optimizes for the privacy and sovereignty of users.
NEAR Foundation commits to incubate and invest in NEAR-based AI infrastructure projects including compute and inference, data curation and rewarding creators, crowdsourcing and synthetic data marketplaces, agentic infrastructure, and more. Any AI application and novel consumer use case can leverage NEAR for the best User-Owned AI infrastructure and access everyone in Web3 via Chain Abstraction.
We have established the @near_ai AI R&D lab led by @AlexSkidanov and me, focusing on building an AI Developer to allow anyone to build end-to-end Web3 applications without code.
NEAR is the best place to build and incubate the next generation of AI projects. NEAR Foundation, as a well-resourced not-for-profit with strong expertise in both AI and Web3, is positioned to drive a User-Owned AI movement that offers alternatives to closed, centralized players and succeeds at a global scale.
Projects building in the Web3 x AI space should apply at: hzn.xyz/hzn
Talk/workshop at @eth_seoul_ Saturday, June 3 at 10am about "Building Composable Decentralized Frontends" – I'm also a hackathon judge. Be sure to check out the prizes from @NEARProtocol@proximityfi and hack on #NEARIsTheBOS!
Talk at @buidl_asia on Tuesday, June 6 at 10:10am: "The Future of Web3 is the Blockchain Operating System." Join for some alpha and learn what's next on the BOS.
Also speaking on an AI+blockchain panel at @buidl_asia on June 6 at 11:25am.
2022 has been a crazy year, let’s recap and sneak preview of 2023!
NEAR Protocol has raised $500m. Startups in the ecosystem raised over $330m. A number of eco funds have started like @MetaWebVC@LyrikVentures@stealthcapfund and many of the existing VCs has been supporting projects on NEAR
I mentioned on Forkast: it’s great time to start now in web3 as early stage capital is there. Valuations will be lower and one must really have the fundamentals down (strong team, good product vision or differentiation in the growing market).
Security should be thought with usability in mind. They are not opposed to each other as some think.
Current self custody is not secure nor usable. Users will go via path of least resistance and expose themself to hacks and phishing if products don't offer usable defense.
Currently web3 space is full of hidden security problems that are usually pushed to the user to DYOR.
Instead let's focus on build products that offer protection by default.
One of the biggest decisions was to make NEAR account model support many keys with different permissions
This allows:
- rotate keys: disable existing key for a new one
- different devices have different keys
- 2FA via multisig with semi-trusted party like wallet service
- delayed recovery via semi-trusted party like bank or group of friends
Blockchain embodies transparency. Consensus is the first layer of mutual agreement on its state. Privacy is the next layer, regardless if it’s via pseudo-anonymity or ZK proofs.
Making someone liable simply for using a privacy tool used also by nefarious actors is equivalent to considering all clients of JP Morgan guilty because someone has used a bank account to launder money.
At the same time, I truly believe that as a web3 community we should make it our goal to counteract Money Laundering and Terrorist Financing. Web3 builders and regulators need to meet in the middle, improving processes and products whilst not impinging on users’ privacy.
From the day 1 with @NEARProtocol it was about experience.
NEAR Ecosystem is ready for mass adoption now. Multiple examples of hiding blockchain and integrating into web2, onboarding of non crypto users and more.
Developer experience was first as the initial team were developers, focusing on bringing common languages (Rust/Typescript), using widely adopted platforms (WASM), scalable infra and programming model that's familiar to web devs (services)
Close second was user experience: remove need to learn about keys, addresses, allow multi device security and progressive onboarding. Making accounts programmable to allow proxy & meta tx.
Making fees predictable (EIP-1559 like) and relying on scalability to keep them low.