Dominic Williams, Creator of DFINTIY, Discusses Decentralized AI, AI Dapps, Blockchain Hosting for AI Models, and Multichain DeFi | Episode 321

Cryptonews Podcast host Matt Zahab recently conducted an exclusive interview with Dominic Williams, the Founder and Chief Scientist of the DFINITY Foundation, a significant contributor to the Internet Computer (ICP) blockchain and a passionate advocate for AI.

Williams elaborated on the prospects of AI, its ‘ideal collaboration’ with smart contracts and blockchain technology, and how AI can assist everyday individuals in executing intricate trades using verbal commands.

He also clarified what sets the Internet Computer apart as a crypto cloud compared to other blockchains, and why it currently occupies a unique position in the “third lane.”

During this interview, Williams covered:

• decentralized AI (DeAI);
• the necessity for AI projects to be decentralized;
• the importance of integrating AI into dapps;
• developing a top-tier content studio;
• hosting AI models on the blockchain;
• multichain DeFi;
• decentralized cloud computing.

Dominic Williams provided a comprehensive exclusive interview, which can be viewed above or read below.

The Future of AI is Smart Contracts

In a new studio located in the basement of Zurich HQ, Williams spoke about the exciting future of AI.

The Internet Computer (ICP) blockchain is designed to operate as a decentralized cloud. This would allow AI models to function entirely on the blockchain without reliance on centralized cloud services like Amazon Web Services.

Williams suggested that “the long tail of AI, where individuals run their own models, is likely where AI executes smart contracts on blockchains.”

This capability is already available on the Internet Computer blockchain.

Currently, discussions in the crypto space revolve around AI: blockchains for AI, tokens for AI, and ChatGPT integration that users can leverage to incorporate chat into their Web3 services, among other topics.

While all of this is positive, Williams emphasized that DFINITY is more concentrated on the foundational aspects.

The project’s goal is to transform computing on blockchain. This indicates that they perceive blockchain as it currently exists on the Internet Computer.

Smart contracts on the Internet Computer consist of a collection of web assembly bytecode and persistent memory pages, operating in parallel. They directly facilitate user experiences in end users’ browsers.

This particular blockchain, according to Williams, can function as a “crypto cloud where you can create anything with smart contracts.”

This enables it to handle approximately 330,000-350,000 Ethereum-equivalent transactions per second.

Additionally, some of these smart contracts are now utilizing up to 400 gigabytes of persistent memory, said Williams, offering sufficient memory and computational power to support an AI.

Earlier this weekend, the Internet Computer #ICP surpassed 3 billion blocks It’s likely now processed significantly more blocks than all other blockchains combined (?) It’s a powerhouse.

— dom.icp ∞ (@dominic_w) March 24, 2024

“What we’ve accomplished at DFINITY,” Williams added, “is the development of technology that enables the Internet Computer blockchain to perform computation at scale. Today, it manages many orders of magnitude more data and computation than conventional blockchains. That’s why it can execute AI smart contracts.”

AI and Blockchain are Perfect Partners

It is well-known that OpenAI has made a significant impact globally. However, according to Williams, it and similar initiatives represent “centralization on steroids.”

Nonetheless, he noted that while these large language models with substantial financial backing attract considerable attention, “in the future, we will witness millions of individual AI models operating.” These will concentrate on specific functions.

In fact, we are already engaging with AIs daily. Common examples include platforms like TikTok or Instagram. Such content recommendation neural networks, for instance, will be highly prevalent in the Web3 domain, Williams observed.

There are several critical aspects to consider, the founder cautioned.

Standalone large language models will coordinate organizations. An example is Delphi AI, which connects to all facets of a technology company, including the calendar, Slack, GitHub – any and all applications.

This implies that the AI model will have access to all of this organization’s most sensitive information, making it a prime target for cybercriminals.

Moreover, a malicious entity could infiltrate the company and provide the AI with false information. The AI cannot distinguish between genuine and fabricated data, leading to corruption.

Consequently, employees within the organization should engage with this type of AI only through robust blockchain-based authentication.

Additionally, the organization would become entirely reliant on the AI. If this AI operates on traditional infrastructure and something trivial goes awry – such as a piece of equipment malfunctioning – and the AI becomes unavailable, it would be “extremely disruptive.”

All of this underscores the necessity for AIs to be tamper-proof. Williams stated that this is accomplished through smart contracts.

When AI is accessed via strong authentication through blockchain, “the AI becomes unstoppable.”

Thus, Williams concluded, “when you consider it, AI and blockchain are ideally suited for one another. Blockchain is the sole means of establishing a tamper-proof, unstoppable computing platform.”

That said, if one seeks the most efficient AI, it should be operated on a dedicated AI cloud, the founder added.

Power of AI-Based DeFi

Blockchain technology is intricate. The realms of crypto, decentralized finance (DeFi), decentralized applications (dapps), and other sectors of the industry can be challenging for the average user to grasp on a technical level.

“Things become complicated when you need to empower individuals to create nearly arbitrary instructions,” Williams contended.

However, this is where AI can play a role. Users can simply articulate their desires.

If a project piques a user’s interest, they can take out their phone and verbally instruct an AI crypto wallet to execute a transaction.

Williams provided another illustration. A user can command the AI to sell 5 WBTC if WBTC on Ethereum reaches $100,000 and only if the slippage is below X.

The AI will then present a straightforward pseudocode – one that is easily understandable – and, if the user approves it, “they press confirm and boom, it’s done.”

Thus, there is also a convenience factor:

“The system can enable ordinary individuals to perform much more complex actions.”

On the Internet Computer, smart contracts can be activated almost instantaneously, Williams noted. They can be periodically awakened to, for instance, check a price oracle and act according to previously established instructions.

This facilitates the buying and selling of crypto and executing trades conveniently, as well as creating intricate default actions based on specific conditions.

Moreover, it is quite simple to develop a basic swap interface.

Witness the groundbreaking fusion of AI and Blockchain that promises to redefine the future

AI running on blockchain as a smart contract pic.twitter.com/Y3pU7gxOEu

— DFINITY (@dfinity) March 26, 2024

While the product is not yet finalized, the initial AI demonstration includes image classification. Users can upload a photo, and the AI will recognize it.

DFINITY is also collaborating with partners and AI specialists, and “they’ve got various initiatives they want to implement on the Internet Computer.”

The project’s concepts include facial recognition, which would facilitate fully autonomous KYC.

Hosting AI Models on Blockchain

Executing AI as a smart contract necessitates a certain degree of replication.

The Internet Computer allows users to select their desired level of replication. On a standard system subnet, it operates with 13x replication, Williams stated.

In the future, there may be lower trust AI subnets, which would only have 4x replication, still providing impressive security and resilience, he argued.

The higher the replication level, the more queries can be processed.

Currently, the Internet Computer’s limitation lies in the amount of main memory allocated to a smart contract: 4GB each. This is due to its operation on 32-bit WebAssembly (wasm).

However, Williams mentioned, “I’m genuinely excited about this. I’m pleased to announce that we’re transitioning to 64-bit wasm soon, which will eliminate that restriction and likely increase the maximum main memory to 32GB or something similar.”

This advancement will enable the blockchain to run large language models. (For clarification, large language models include GPT-3 and GPT-4 from OpenAI, LLaMA from Meta, and PaLM2 from Google.)

Meanwhile, DFINITY’s development pipeline already incorporates Single Instruction/Multiple Data (SIMD) instructions. These will be integrated into the smart contract execution environment to “significantly” accelerate these AI smart contracts.

“And we have numerous other initiatives in progress as well,” Williams stated. Initially, they plan to host various AI demonstrations.

Furthermore, they are working on two major projects: one is Utopia, and the other is Orbit – an open-source initiative that will allow users to establish a fully decentralized, institutional-grade, cross-chain crypto custody solution on the Internet Computer.

Creating Utopia

Regarding future developments, Williams highlighted two primary areas of focus for the team.

One is optimizing the Internet Computer to support AI so it can accommodate larger models and do so more efficiently.

The other is Utopia, which will enable the enterprise sector, governments, NGOs, and others to operate like private Internet Computers that can integrate with the public Internet Computer and with one another.

Individuals utilizing Utopia to establish these private clouds will “simply concentrate on ensuring that online systems and services are tamper-proof and unstoppable.”

Another crucial point to consider is that the team behind the Internet Computer is consistently achieving advancements and milestones – yet there is no “fanfare.”

On other blockchains, trivial features are often “wrapped up in a grand release with an elaborate name, and then heavily marketed.”

However, the Internet Computer “isn’t the best” at self-promotion. Williams noted that it tends to be overly focused on technical aspects.

This is also an area they are working to improve.

An excellent example of this is a concept called deterministic time slicing. It’s a significant accomplishment, Williams stated. It essentially demonstrates that the Internet Computer is a legitimate, decentralized operating system.

While operating systems like Google’s Android, Apple’s iOS, OSX, Windows, or Linux function on a single computer, the ICP stack “operates as a type of operating system that runs across multiple computing units.”

These “developments are far more advanced than what people are accustomed to on blockchain,” Williams remarked.

The Three Lanes of Blockchain Technology

Finally, in relation to the previous discussion, Williams pointed out that there are three lanes concerning blockchains.

The first lane consists of blockchains dedicated to maintaining a ledger.

The second lane is for blockchains that support smart contracts.

The third lane encompasses “the world computer blockchain.” This is where the Internet Computer resides, and it is challenging to convey that, according to Williams.

#ICP’s technology is years ahead, and will remain so.

“A blockchain that supports full stack decentralization, where you can build social networks entirely on-chain, and even run AI as smart contracts.” – @dominic_whttps://t.co/G6JFBDGvKB

— DFINITY (@dfinity) March 28, 2024

People often assume it belongs in lane two, viewing it as one of the conventional blockchains that host smart contracts with “less computing capacity than a Raspberry Pi.”

However, this blockchain is already a serverless cloud where users can create online systems and services utilizing smart contract software.

This is distinct from what leading blockchains offer, Williams stated. The difference is significant to communicate.

Blockchains typically store “very small amounts of data and perform very limited computing. […] The gap between the current state of blockchain and its potential to evolve into a crypto cloud is substantial,” Williams concluded.

__________

About Dominic Williams

Dominic Williams is the Founder and Chief Scientist of the DFINITY Foundation, a key contributor to the Internet Computer (ICP) blockchain.

He is a theorist in crypto and an entrepreneur. He has been active in the blockchain sector since 2013. Prior to that, he was an engineering entrepreneur who developed multiple internet technologies and products.

Williams specializes in distributed computing and crypto network theory, having proposed numerous innovations that are currently in use.

The post Dominic Williams, Founder of DFINITY, on Decentralized AI, AI Dapps, Hosting AI Models on the Blockchain, and Multichain DeFi | Ep. 321 appeared first on Cryptonews.