The Web3 and AI Newsletter 13
Hello, everyone, and warm welcome to all newly joined subscribers!
This is the Web3 + AI newsletter, where we explore the intersection of blockchain and artificial intelligence. If you follow the so-called Crypto Twitter, you should be aware that we are in a state of a full Web3 + AI craze! Every day, there are new projects coming out, major deals signed, and innovative use cases revealed. Influencers are racing to publish the latest Web3 + AI report, and traders are fueling speculation. Happily, you have me to disperse the noise and collect the most interesting updates consisely listed below!
Thank you for being here! Let's dive in!
What's New in the Web3 + AI Startup World?
Polkadot Releases PolkaBotAI to the X Users
Polkadot released PolkaBotAI - an education AI assistant, empowering X users to learn everything there is to know about the Polkadot chain and the DOT ecosystem. All one has to do is tweet: “/ask @PolkaBotAI” and pose a question.
PolkaBotAI is built in cooperation with OriginTrail, guaranteeing information provenance, and NeuroWeb.ai, which is responsible for the knowledge mining throughout X.
Optimism to Use EZKL to Secure RetroPGF Initiative
Optimism, Ethereum's prominent L2 chain, will be employing EZKL's zero-knowledge proofs to verify and secure its Retroactive Public Goods Funding (RetroPGF) initiative. Ever since 2021, Optimism has been rewarding community projects, influencers, and contributors for their work in advancing the chain's adoption. However, with the protocol's skyrocketing popularity, it is becoming difficult to guarantee that the RetroPGF grants are allocated according to the pre-set rules. This is where EZKL comes in to help.
Such a guarantee involves a proof that votes were legitimate and that the allocation algorithm was run as intended, whilst keeping the votes underpinning the RetroPGF allocation private. Which is exactly the sort of problem zero-knowledge (ZK) cryptography was born to address.
Why EZKL? Optimism's allocation algorithm was written in Python, and that EZKL makes it easy to convert many Python data science calcs into ZK proofs, our stack fit like a glove.
HyperGPT Partners with Web3Go
HyperGPT, a Web3 marketplace for AI solutions, has struck a partnership with the data intelligence network Web3Go.
This partnership is designed to leverage the strengths of both platforms, integrating HyperGPT’s innovative AI solutions with Web3Go’s advanced data intelligence network. Together, we’re pushing the boundaries of what’s possible in the Web3 space, enhancing the development, accessibility, and utility of AI-driven applications and services.
The financial dimensions of the deal were not revealed. The HyperGPT store lists various AI models and provides a seamless SDK for integrating them, whereas Web3Go collects consumer data and uses AI to source actionable insights.
HyperGPT
A recent X Spaces session by Delysium hosted both HyperGPT and Web3Go and painted a richer picture of what they are doing in Web3 + AI. Listen to it here, but remember - X spaces recordings last for 30 days only.
Alaya Allies with MUA DAO, Shareverse, and CyberPetaCraft
Over the last few weeks, the distributed AI data platform Alaya has announced a series of interesting partnerships. First, it joined forces with Metaverse Universal Assets or MUA DAO to boost the AI capabilities of its MUAVerse.
MUAverse encompasses an AI-driven interactive engine and a multi-Metaverse platform. It showcases the product of MUA as a Metaverse middleware infrastructure. Whether you are a player or a business entity, all your assets will be presented and traversed in the most engaging manner across multiple chains and diverse art styles within the Metaverse.
Later on, Alaya partnered with Shareverse, another AI-powered metaverse platform for all builders, artists, and creators alike, and with CyberPetaCraft - a tool for gamified AI agents training through Social Entertainment on Twitter. No additional details about any of these arrangements have been shared yet.
Compute DePINs Power Metaverse Experiences & AI Training
Two recent news stories exemplified the value of computation DePINs and highlighted why they are increasingly preferred to legacy cloud services like AWS or GCP.
First, the virtual experiences platform Passage launched its first metaverse world powered by Akash Network's decentralized and permissionless GPUs. Passage testifies that the advantages of the switch are threefold: reducing costs, enhancing scalability, and preserving sovereignty and privacy for creators.
The results from the first Passage Worlds hosted on Akash have resulted in compute savings up to 70% cheaper than AWS or Google Cloud with comparable hardware (high-density P100s and RTX 3090s).
Moreover, Nuklai, a collaborative data marketplace aimed at facilitating AI training, gave its users access to GPU.net's peer-to-peer compute infrastructure to provide an alternative to steep personal GPUs.
Our users will have more computing power choices, creating a competitive environment that benefits everyone. This flexibility helps our community to optimize their projects for efficiency, cost, and performance. It gives them freedom of choice, further democratizing access to computational resources.
Unlockd Adds AI Price Predictions by Allora
Allora, the decentralized AI network I told you about a couple of weeks ago, has announced the integration of its AI price predictions into the RWA (real-world assets) liquidity protocol Unlockd.
Now, leveraging Allora’s AI price predictions, Unlockd offers peer-to-pool loans in its V2 testnet, allowing instant access to funds against RWA NFTs. This integration sets a new standard for instant, permissionless liquidity.
Unlockd has chosen Allora in particular because of its proven record in pricing various digital goods like NFTs or real-world assets with 95-99% accuracy, its tamper-proof outputs using ZKML, and its permissionless and universal access to AI. Learn more below.
Greg | Upshot
Giza Product Demo for Gnosis
The February Gnosis AI call featured Giza presenting its decentralized AI stack and emphasizing its mission to build a universal technology that is suitable for both enterprise and personal needs.
The Giza team showcased their open-source ZKML framework Orion which you can use to compile your ML models into ZKML circuits so that you can run verifiable inference on them. They also demoed the Actions SDK that “empowers developers to build and scale Verifiable ML solutions quickly, turning their python scripts and ML models into resilient, recurrent workflows”, and the Giza Datasets - the Hugging Face for Web3 ZKML use cases. Check the recording below to learn more.
First AI-Related Grant by the Ethereum Foundation
The Ethereum Foundation Ecosystem Support Program awarded Kaito AI a grant, marking the first case for an AI-related project to receive support. Kaito AI will use the money to democratize access to Ethereum-dedicated real-life events. In practical terms, Kaito AI intends to employ AI, and NLP (natural language processing) in particular, to transcribe hundreds of conference talks held at events like EthDenver or EthCC and distill valuable insights.
Kaito AI is on track to building THE AI super app in crypto, from a search engine to a smart recommendation system, news firehouse to advanced social analytics, and expanding from serving institutions to ultimately serving everyone in the space.
Noteworthy Podcasts, Articles, Events
ETHSofia Conference & Hackathon 2024
And since I mentioned Ethereum real-life events, allow me to invite you to the inaugural ETHSofia Conference and Hackathon taking place in Sofia, Bulgaria in October. I am happy to share that I am part of ETHSofia's founding team and that we are determined to create a world-class event gathering hundreds of cypherpunk-inspired entrepreneurs, hackers, and builders.
Over the coming months, we will be sharing more information about partners, speakers, and side events. If you want to take part in any way, don't hesitate to reach out to me!
Web3 + AI on Solana X Spaces
I usually avoid sharing X Spaces here, because my mission for the Web3 + AI blog is to be a permanent repository of useful resources you can always come back to consult. However, this one X Spaces session on AI + Web3 on Solana was quite interesting, and you still have about a month to listen to it.
It featured several Web3 + AI projects building on the Solana blockchain such as Grass, Synesis One, Io.net, EZKL, as well as other relevant protocols like Gensyn. They held a fruitful discussion on the various types of challenges Web3 + AI projects face, including data authentication, payments, management of decentralized computation resources, and proofs of personhood. They even explored the possibility of custom AI chips, similar to Bitcoin's ASICs, to appear. Check the recording out and let me know what you think about it!
Disclaimer: None of this should or could be considered financial advice. You should not take my words for granted, rather, do your own research (DYOR) and share your thoughts to create a fruitful discussion.
Member discussion