Fetch.ai AMA Recap

Published in

Citadel.One

9 min readMar 18, 2022

Fetch.ai is an AI lab building an open access, tokenized machine learning network to enable smart infrastructure built around a decentralized digital economy.

On March 16th, 2022, members of the Fetch.ai team joined us for an AMA, to answer some questions from the Citadel.one team and the crypto community.

AMA participants list
1 Stage. A set of questions prepared by the Citadel.one team.
2 Stage. Prior to the AMA, Telegram and Twitter users asked their questions for the Fetch.ai and Citadel.one teams. The best 5 were chosen and answered.
3 Stage. The chat was opened for the community to ask more questions about Fetch.ai.

AMA Participants

Guests

Atari, Fetch.ai Community Manager
RV, Fetch.ai Marketing Operations Manager
Niko, Fetch.ai Ambassador / CM
Rina Shtedle, Citadel.one CMO & BD

Host

Vlad Gorlov, Citadel.one Head of Community & PR

STAGE 1. Questions from the Citadel.one Team

1. Can you explain what Fetch.ai is about in simple words? How are you going to implement artificial intelligence and big data in your product?

RV: Fetch.ai is in essence built around Digital Twins. Formerly known as Autonomous Economic Agents. We firmly believe in the idea of removing middle-man or aggregator in today’s digital economy. Which is where our Digital twins can play a crucial role.

Atari: The Fetch.AI network incorporates machine learning (ML) and artificial intelligence (AI) capabilities with blockchain tech via digital twins — also known as Autonomous Economic Agents. These agents connect the dots between vast data sources that can be used to enable products spanning data-intensive industries like supply chain, healthcare, travel, and DeFi.

It’s all part of what’s known as a “multi-agent system” which is a field in AI that most people associate with having a large amount of intelligent “agents” looking to solve complex problems that could be considered very difficult, or near impossible to solve by any one person. So instead, these agents communicate with one another to learn the best possible solution to many difficult problems.

Niko:
Digital twins can execute meaningful economic work such as buying/selling/monitoring/arbitrage/negotiating and more and are particularly useful in complicated industries where the coordination and cooperation between multiple parties are paramount. In order to achieve this, the team has developed a smart ledger an open economic framework an agent communication network and built multiple libraries as well as a virtual machine including multiple other components which you can explore more in-depth here.

2. Tell us about the background of the Fetch.ai team. Who is behind it? What are your roles and tasks there?

RV: Well, you’ve Humayun Sheikh who is the CEO and Founder, Jonathan Ward who is the CTO, Kamal Ved who is the Chief Product Officer on the fetch foundation side of things, Devon Bleibtrey who is the Chief Product Officer on the Digital Twin applications and Maria Minaricova who is the Director of Business Development.

We are in total a team of 70+ spread across the globe including contractors. Back when Fetch.ai was conceived as an idea, we’d Humayun, Toby (former COO but advisor now) and Thomas Hain who were at the helm.

The company has scaled considerably since then. On our community side, we have Niko and Atari here who are our community managers/admins along with Trema and many many more.

My role is pretty much handling all the operations on the community side and also marketing/PR etc. Atari and Niko are in-house tech community experts when the community has any question either on Cosmos or anything related to Fetch.

3. We saw your awesome video about the use-cases of Fetch.ai technology. Could you please briefly introduce some of them to our users and followers?

RV: We needed a new video that would demonstrate the vision and the goals we have as a company. We began working on the video back in December last year knowing we want to have it ready in early 2022. So we packed in all the goodies including our use-cases and it turned out to be picture perfect.

Bringing Fetch.ai’s vision to life through videos is the toughest thing to do and we are not joking at all. In today’s world, you get just about a minute to grab someone’s attention and if you don’t get them within 60 seconds, you can forget it. So the video had to convey the vision of Fetch.ai. What is the vision you might ask? Well, AI for everyone. That’s it. In essence, we don’t want AI to be thought of as a Minority Report concept or Skynet. AI must work for you. It should not scare you. It should revolutionize today’s economy. Give everyone a fair chance. If we can make someone’s life better using our digital twin — for example an uber driver who can stand to earn more using our technology than under Uber, we will take that chance.

4. The Fetch.ai blockchain is an interchain protocol, based on the Cosmos-SDK. You already mentioned why you chose to be a part of the Cosmos Ecosystem and that it’s a perfect match. Now, according to your imagination, what potential do you see in Cosmos?

RV: Basically, we are an interchain protocol, based on the Cosmos-SDK, and use a high-performance WASM-based smart contract language (Cosmwasm) to allow advanced cryptography and machine learning logic to be implemented on-chain. This also allows the Fetch.ai network to also serve as a layer-1 network for legacy blockchain networks and as an interchain bridge to the outside world. So if you’re an EVM project looking to move to Cosmos, we can help you. If you’re a Cosmos project looking for on-chain analytics, we can help you. There is something specific to fetch.ai which are our AEAs/Digital Twins and we want to see them everywhere. There’s no other chain like Cosmos that would make it possible. Simple as that.

Fetch.ai on Cosmos = Scalability + Interoperability.

Niko: In short, the cosmos ecosystem allows for unprecedented horizontal scaling and integration. Allowing blockchains to benefit from each other’s strongest value offering. The goal within the cosmos ecosystem is to allow each interested chain to explore and apply automation, analytics, autonomy in a decentralized manner.

STAGE 2. Questions from Users

1. AI is supposed to learn based on the data it received, are you sure that current blockchain capacities are enough for real artificial intelligence to step forward? Because we are talking big data here, I assume it’s lots and lots of information. Aren’t you afraid the network will simply halt once AI becomes too smart? Cause right now even an airdrop can cause a network to fall.

(by @katusuna)

RV: This is a fairly technical question and Katusuna can find more info via our smart ledger paper. BUT we can do it in a bit of detail here.

The fetch.ai ledger is designed to support huge transaction volumes in practice. This means it must scale to millions of transactions while still delivering rapid confirmation times. The blocks in the Fetch ledger are divided into a fixed number of slices that can each contain many transactions.

I am sure Niko can dive in more but I do want to leave Katusuna with our whitepaper on Ledger which will be immensely helpful.

Niko: Although there are computational and temporal constraints, not all transactions need to be on-chain transactions. The Fetch.ai ledger uses a combination of DAG technology (to allow for fast off-chain training) and a novel resource optimization model that groups transactions into lanes allowing for parallelism whilst maintaining the order of each transaction.

2. Beyond the Collective Learning approach, are you planning by any chance to incorporate some sort of evolutionary logic that could enable AEAs to acquire special traits in response to changing market conditions, which could eventually improve the decision-making of an AEA?

Niko: Depending on the purpose of the AEA its decision-making will continuously improve as it encounters more data and refines its ability to decide (optimize against some set of criteria). The acquisition of skills is already possible.

3. What is the level of privacy of the fetch.ai blockchain? Can validators see the computations or is it run inside trusted execution environments?

(by @Raimo099)

Niko: Privacy when training a model is ensured through a method called Epsilon-Differential privacy. In short, this means randomizing and encrypting the data before anyone has any “access” to it. You can find a specific answer to the definition of privacy, the methods used to ensure it, and how it’s instantiated in the CLI here.

4. Already 8 weeks past for CitadelDAO being a validator on Fetch ai so far. What are the preliminary results that both parties had? Could you share some insights?

Niko: On our end, we’ve gained a trusted, secure, stable, and engaged validator which by default ensures the decentralization and security of the network. Given that this is the core premise of validating — we’re happy that the Citadel team showed interest in our project and onboarded themselves as a validator.

Atari: Well I would say on a base level Citadel.one already has a well-established reputation within the Cosmos ecosystem and so they’ve been able to help us reach further into it as we attempt to make ourselves known. They’ve even integrated us into their wallet application for staking (and soon with their external rewards as well) and I’m sure our partnership will continue forward with including Fetch applications that are accessible in the wallet application as well. As a validator, they’ve done a great job and been extremely stable during the time they’ve been up. So all in all it has been great.

RV: We are quite stoked to have Citadel on our chain as a validator. Their experience across the Cosmos ecosystem speaks for itself and we want to bring onboard validators that the community can trust and delegate their stake to. They are highly engaged and share content like weekly recaps, organizing this AMA as well as willing to support us with services that our community can use — namely staking/wallet. Obviously, we are keen to deepen this partnership and that’s what we plan to work towards.

Reena: So far we happily integrated Fetch.ai into our platform enabling all the asset management functionality for FET holders. You can read more about it here.

Soon we will be introducing XCT rewards for staking with Citadel.one node, so please stay tuned with our official channel.

We are also considering adding Fetch to our Governance extension :)

In the future, we are looking forward to integrating applications and use-cases of Fetch to strengthen both ecosystems! We will be happy to welcome outside developers very soon!

STAGE 3. Live Chat

You can read the full live discussion in our Telegram chat, starting with this message.

Niko: if your question was left unanswered or you want to read more feel free to check out our other most recent AMA on the cryptocurrencies subreddit with our CTO and our Chief Product Officer here.

Stay tuned for future activities: follow Citadel.one on Twitter and Telegram.

If you have further questions, drop us a message to support@citadel.one. You can also join our chat in telegram, there is an active discussion of all current updates, as well as round-the-clock support from our managers.

About Citadel.one

Citadel.one is a non-custodial Proof-of-Stake platform for the management and storage of crypto assets. Users can create public addresses for all supported networks with one seed phrase, connect their Ledger or Trezor device, or import an address generated by another wallet.

The analytical dashboard provides relevant information on wallets’ balances and networks’ main metrics. In Citadel.one, we standardize the semantics, making interactions with the platform as easy as possible. The same goes for network metrics. By using universal terms, we are making it easier for users to understand and compare networks.

One of the main functions of the Citadel.one platform is participation in the PoS consensus — users can stake and delegate their assets, claim rewards, and follow the latest network proposals in the voting tab.

Citadel.one offers its users instant cryptocurrency exchange services that allow fast and secure crypto assets swap. It is also possible to buy and sell crypto with a credit or debit card. Citadel.one users can track rewards, withdrawals, transfers, and deposits across all supported networks and add comments to these transactions.

Among PoS platforms, Citadel.one supports BAND network ($BAND) Secret Network ($SCRT), Cosmos ($ATOM), ICON ($ICX), IOST ($IOST), KAVA ($KAVA), BSC ($BNB), Juno ($JUNO), Comdex ($CMDX) and Tezos ($XTZ). We also support Ethereum, Bitcoin, and Tether ($USDT) for our users’ convenience. Mobile and desktop versions, new networks, including Polkadot are scheduled for the upcoming updates. Furthermore, delving into the idea of true decentralization, we envision Citadel.one a decentralized autonomous organization and a genuinely community-owned platform in the nearest future.