Discover - Article 1
Long developed in parallel, AI and Web3 now seem on the verge of converging
Discover - Article 2
DePINs for decentralised AI management
Discover - Article 3
Blockchain and AI: Synergies across the entire AI value chain
Discover - Article 4
Jonathan Schemoul (Aleph.im): "True AI with free will can only be achieved with decentralised technologies".
Go deeper - Article 5
How blockchain can help you prove your humanity in cyberspace
Go deeper - Article 6
Worldcoin, the flagship AI project of the crypto
Go deeper - Article 7
When crypto miners diversify into AI
Go deeper - Article 8
Map of Web3 players working in AI
Go deeper - Article 9
Market analysis: how are AI tokens performing?
Go deeper - Article 10
How AI can help make blockchain more secure
No items found.

Jonathan Schemoul (Aleph.im): "True AI with free will can only be achieved with decentralised technologies".

Subscribe to our smart subscription to read this report in full

Jonathan Schemoul (Aleph.im): "True AI with free will can only be achieved with decentralised technologies".

The Aleph.im network offers a decentralised infrastructure for developing AI tools that are less dependent on the Web giants. We met its creator.

Tout ce qui compte en Web3. Chaque semaine.
25€/month
Try insider for free for 30 days.
Subscribe

What is Aleph?

Aleph.im was created in 2019, with the aim of filling a need: that of having a decentralised cloud, because smart contracts are not enough if we want to create complete applications. Last January, we launched Twentysix Cloud with the idea of separating the infrastructure part of our various solutions for the public and businesses. Twentysix Cloud brings together our computing, storage, VRF, indexing and custom domains offerings, while Aleph manages the infrastructure with all the independent players who manage the various cloud nodes.

How is your proposal more relevant than a centralised option?

We offer much greater resilience: if a datacentre goes down or a bug occurs, the system redirects requests. What's more, if one of the providers is refused hosting, the content/computing power also moves. This provides de facto resistance to censorship. Of course, for really problematic content (illegal content, for example), no hosting provider will agree. Self-hosting remains a solution in this case (with the risks that this entails).

Is it competitive in terms of price?

Actually, in pay-as-you-go, the price per hour is equivalent to that of AWS. Once pricing is activated on the network (by the nodes themselves) this will be even more the case, with increased competition between players. This will make it possible to achieve lower prices than centralised competition.

How does Aleph's internal economy work and how does your token fit into it?

We have grouped all our solutions under Twentysix Cloud in order to differentiate them from the infrastructure part (aleph.im). It is currently possible to pay for the various products in 2 ways:  either by holding ALEPH in your wallet, or via our Pay-As-You-Go system. Payment is made in crypto via Superfluid on the Avalanche blockchain at millisecond intervals. We plan to extend our Pay-As-You-Go system to stablecoins as well as fiat.

Staking keeps the network decentralised. For this Core Channel Node managers must deposit 200,000 ALEPH on their wallet to start their node. Once 500,000 ALEPH have been staked on a CCN, it provides its delegators with a staking reward of around 5%. Staking starts from 10,000 ALEPH, on Ethereum and involves no transaction fees or vesting period.

With regard to Compute Resource Nodes, the nodes that provide computing power to the cloud, these are remunerated between 1,200 and 1,500 ALEPH per month depending on their decentralisation quotient, stability and power. We held a halving on 21 March on the rewards distributed to CRN holders.

Can you give some figures on your business?

We are approaching 100 Core Channel Nodes, i.e. 100 nodes that help manage and maintain the cloud. The more CCNs we have, the more decentralised it becomes. At the same time, we are seeing an increase in the number of Compute Resource Nodes, the nodes that will provide computing power to the cloud and enable us to deliver ever more powerful solutions. There are currently 104 million tokens keeping the cloud up and running, including 85 million in staking and 19 million by Core Channel Node holders out of 247 million tokens in circulation.

Who are the best-known members of your network?

Currently, our main partner is Ubisoft, which we are supporting on a number of infrastructure issues in Web3 gaming and which has been managing its own Core Channel Node since 2021. We have also accelerated our development with students through the Epitech and Kryptosphere POC associations, bringing together a number of prestigious schools and universities to help students learn about the new technologies in our cloud. 

Our aim is also to enable businesses to participate in the decentralisation of the cloud by managing their own nodes, as was recently the case with ACI, which is one of the most important service providers in the Aave DAO. We work closely with several blockchains such as Avalanche, BNB Chain and Solana in order to port our solutions to their networks.

What types of services do you offer projects evolving in AI?

Mainly the launch of serverless models on the network. This allows them to scale easily. For those who don't want to bother managing their models, we have APIs via Libertai on several models. This API is compatible with the OpenAI API. Libertai is also working on a framework for deploying autonomous agents (which can be given a task and will work autonomously) that will be launched as virtual machines on the network.

Why is it relevant to decentralise the way AI works?

When it comes to AI, decentralising it fits in with the same logic as DePIN in general: Offering a robust, censorship-resistant and agnostic AI service. There is always a visible contradiction between wanting to build DeFi/GameFi/SocialFi dApps, and hosting them on AWS/Google Cloud or with centralised dependencies (offchain storage and proprietary AI, for example). Genuinely decentralised, open-source AI has the advantage of being permission-free, bias-free (no prescriptive fine-tuning) and data private (no collection without permission). Web2 and Web3 applications can rely on AI with a decentralised infrastructure to take advantage of their products without being held hostage (and their users with them) by highly opaque proprietary solutions.

There is also a need for AI that is still fairly poorly understood on the blockchain side. It can be used to simplify the user experience on the application side, facilitate data processing and delivery (indexing, data mining, formatting, analysis), or automate off/on-chain processes (the use of agents for DAOs, oracles, etc), and content generation in GameFi/SocialFi too.

What AI/Web3 convergence do you anticipate in the next 5 years?

The convergence of AI and Web3 allows AIs to self-replicate and pay for their own computing. By AI I mean here independent agents. A still under-exploited part of this convergence is that with DAOs: a set of agents will be able to form a DAO that pays for its own computing power, in a decentralised way and without a human intermediary.

When we think of artificial intelligences, we imagine beings endowed with free will with great capabilities. These are the AGIs (general artificial intelligences) or AGIs in English. We're still a long way from that today, but I think that a true AGI will only emerge with the decentralised technologies of web3.

Previous

Next