Fetch.ai CEO Humayun Sheikh participates in AMA with Binance

Oct 3, 2019

On Friday, Fetch.ai chief executive Humayun Sheikh took part in an AMA on Binance’s English Telegram channel. The AMA was hugely popular and lasted more than two hours. In case you were not able to make it, we’ve included a transcript of some of the highlights.

The AMA was broken down into three segments. During the first segment the host of the AMA introduced Humayun and asked him six questions:

Hi Humayun, it is our pleasure to have you today with us and thank you for taking the time to be here. Before we kick off, would you mind telling us a bit about yourself and Fetch.ai?
I am trained as a computer engineer but have spent my last 15 years in trading commodities, developing ML/AI algorithms for market trading and working with various AI projects to deploy commercially valuable usage to AI. My most notable involvement was in DeepMind, which was ultimately sold to Google and is one of the world’s leading organizations in Artificial General Intelligence. AI is fantastic but the problem is that not many people know how to deploy, commercialize and monetize it.

Fetch.ai was born out of the problem of how do you enable everyone to utilize this wonderful and useful technology? How do you connect various things and assets together? And how can you deploy intelligent solutions? Fetch.ai came out of stealth mode in March 2018. We are based in Cambridge in the United Kingdom and our team consists of 64 people, of which the majority are engineers and researchers.

What makes Fetch.ai’s technology unique?
Fetch.ai enables intelligent connectivity. By using a collective learning AI/ML architecture, we enable devices and software agents to find each other, learn from each other and exchange value with each other.

Our core value proposition is the ability to connect devices and enable autonomous actions to deliver solutions to complex tasks. This means you can connect your device with a click and deploy a revenue model around it. For example, we are using our technology to connect all smart city infrastructure, derive intelligence from each connected edge device, carry out multi-party computation and deliver improved journey optimization as a seamless system. The underlying architecture delivers connectivity on different dimensions like control signals, deep learning and economic exchange.

Fetch.ai’s beta mainnet release is due at the end of the month. Is it on schedule and what features will it include?
Yes, indeed we are on time — we pride ourselves on delivering on time. The release at the end of September is called Aquila. As with all of the milestones featured in our roadmap published in February, the release will be delivered on time. More details will be announced on Monday, but the release will include updates to our Synergetic Contracts (complex problem solving). It will also include a significant update to our random beacon implementation as part of our Proof-of-Stake with Unpermissioned Delegation consensus mechanism. This one is a biggie!

Postscript: As promised, our beta mainnet was launched on Monday 30 September. You can read all about it here.

Lots of people in the community are talking about staking. Do you have an update for us?
It is coming ? We will be announcing exciting new details about staking next week (week commencing 30 September). This will include guides as to how to take part in our staking auction both directly or as a member of a pool.

The source code to the auction contract is available to be explored. We are very proud that our staking contract has been awarded a score of 100% for security and implementation quality by CertiK. Watch this space, you will see interesting things happen over the next week!

Fetch.ai’s technology has the potential to disrupt almost every industry. Which sectors are you currently focusing most of your attention on?Although we are sector agnostic, our initial focus is on deploying where the resource and usage is digital and does not require too much alteration to the infrastructure. Our key focus is in the energy sector where there is a high momentum to deploy an intelligent grid system and Fetch.ai infrastructure suites this really well.

Smart cities are a key focus for us too. There are several partners we are working with to build and deliver a solution.

One of my personal favorites is physical commodity trades where the likes of Corda are struggling a bit. Fetch.ai infrastructure finally enables the trade finance and is an easy win, but perhaps not as sexy!

Postscript: On Tuesday 2 October, we announced a partnership with Turkish steelmakers to develop the first AI-powered decentralized metals exchange.

Implementing robust regulation is vital to encourage the widespread adoption of AI and blockchain technology. What efforts are Fetch.AI taking to shape this regulation?
Absolutely! But we do understand that regulation is generally behind the technology curve. So our approach is to work with the regulators and provide them with more understanding on what we are building and for us to understand what the regulators want.

Our CTO Toby Simpson was at the recent Organisation for Economic Co-operation and Development (OECD) Global Blockchain Policy Forum. The forum provides a platform to explore important issues such as the taxation of digital financial assets, sustainable infrastructure investment and the policy implications of blockchain and distributed ledger technology.

As a founding member of Blockchain for Europe (alongside Cardano, NEM, and Ripple), in September we attended a series of meetings in Finland and the Baltic region to discuss blockchain and crypto assets with various governing bodies. We remember that regulators are our friends — let’s not make them our enemies.

In the second segment Humayun asked five quiz questions to the community. A total of 24,000 FET tokens were available for each question, with the prizes being distributed equally among the individuals who answered that question correctly. Overall 120,000 FET tokens (24,000*5) will be distributed to people who gave correct answers in Segment 2. For reference here are the questions and answers:

Q1: Fetch.ai has teams located around the world, but where is the company’s headquarters?
A: Cambridge (UK also accepted)

Q2: What’s the name of Fetch.ai’s unique smart contract language?
A: Etch

Q3: Earlier this year Fetch.ai collaborated with Bosch, T-Labs and Share&Charge to win the Trusted IoT Alliance Smart E-Mobility Challenge. What was the name of the winning submission?
A: My Easy Charge

Q4: Which highly respected smart contract auditor verified Fetch.ai’s staking contract, awarding it a score of 100% for security and implementation quality?
A: CertiK

Q5: What is the name of our most recent technical release?
Aquila (Aquarius also accepted)

In the third segment of the AMA, we welcomed questions from the community. In each round, everyone had three minutes to submit their questions. Humayun then chose the best questions and answered as many as he could during the following 10 minutes. There were three rounds of these question and answer sessions. We’ve included some of the best questions below:

You have mentioned your involvement in DeepMind. What was your official title at the company? Also, can I meet your team when I am in Cambridge?
I was a commercialization lead for two years and our CTO Toby Simpson was head of software development at the company for three years. You are welcome to meet our team in Cambridge. Please connect with us via Telegram to organise a visit!

Many people misunderstand the concept between token price and value. A fall in the value of FET doesn’t mean its value has dropped. Can you explain this point clearly to the wider community? I believe in the company’s long-term vision.
The price of the token is not in our control, but the development of technology and the delivery of our objectives is. The value of the Fetch.ai network is increasing every day as we deliver our technology and related use cases.

IoT technology will require a lot of privacy for day-to-day hardware. Can privacy features be implemented using Fetch.ai?
Yes, privacy features are implemented in Fetch.ai. We live in a GDPR world and privacy is very important. Firstly, our autonomous agents only represent what they have — so the actual data isn’t on the network. When agents are introduced to each other, they transact directly and can do so in an encrypted form. Furthermore, there are other technologies such as privacy preserving computing/multi-party computing/homomorphic encryption that allow people to work on data or execute algorithms without knowing what the data is.

Does the team at Fetch.ai intend to create dapps that integrate artificial intelligence so that users can conveniently use the network?
Yes, we do. AI is at the heart of Fetch.ai. Not only does the protocol use it fundamentally (for trust, reputation and learning how to connect agents together), but it enables users of the network to find and deliver AI. On top of this, synergetic smart contracts allow miners to contribute to grander optimization problems and be rewarded for doing so.

Fetch.ai agents could potentially be used for trading and an example would be that of crypto market makers. But would it be feasible to use an agent for exchange-performing statistical arbitrage and communicating with each other for interexchange arbitration?

Are you working on delivering cross-chain interaction for Fetch.ai?
We are working on cross-chain communications as fast as we can. This will enable users of different chains to use our ML/AI-based contracts and tools. Arbot is a very relevant recent example of this.

There are three core issues prevalent in crypto and blockchain: security, interoperability, and scalability. How does Fetch.ai plan to overcome these issues?
The grand trilemma. It has been said that you can pick two of those, but in reality, you can have all three. Scalability is provided by our smart ledger and our unique proof-of-stake with unpermissioned delegation (our minimum agency consensus). Security is delivered across many levels, from the carefully crafted C++, through the architecture’s design, to the philosophy of incentivizing decentralization. This last one is important: larger populations of simpler things are more robust and secure. Interoperability is something that’s important to all of us in this space and we do it using agents that interface with other projects. We are also looking at other exciting ways of being able to harness other technologies inside Fetch.ai.

Users may not really care whether the underlying infrastructure is decentralized. What will motivate them to use Fetch.ai instead of another platform?
You’re right: users generally don’t care, even though they should. As Douglas Adams wrote: “We are stuck with technology when what we really want is just stuff that works”. If this technology makes your life easier, then it’s a win. And decentralization does do that: it returns control and utility to you, and with Fetch.ai, it pushes solutions to your problems at you. So our focus, as is the case with so many others in this space, is to deliver technology that makes a concrete, obvious difference to your quality of life.

Big data storage often causes issues for IoT technology. How does Fetch.ai address these problems?
Instead of taking data and learning in retrospect, we learn on the fly and learn collectively. This technology allows us to sidestep big data issues.

Only a few thousand of the millions of developers worldwide are using blockchain technology. How do you plan to attract developers to Fetch.ai who are not currently working in this space?
We are attracting ML and AI developers at a very high rate and forming collaborations with various prestigious institutions. Our advisors include senior professors at Oxford University, Cambridge University and Warwick Business School.

How will you solve the transaction conflict problem of packed transactions from different blocks?
Assuming you mean how one resolves the issues over cross-shard transaction packing, then we have a cracking paper that talks about this in more detail which I strongly recommend. Being able to pack-and-stack transactions across many lanesso that we fit as many into a block as possible is a big part of the performance achievement that we get: it’s like a giant Tetris game of transactions. If, on the other hand, you mean issues over asynchronous cross-shard schemes, then that doesn’t apply to us: our blockchain is based upon a synchronous scheme so we don’t have multiple conflicting blocks.

How do you plan to attract more business partnerships to your project?
We are doing this by showcasing our technology stack — and we are getting a lot of traction!

Mainnet is the production ready network, with our native token. It allows users of the network to build agents and build Synergetic Contracts or smart contracts. The network will introduce agents to each other using various AI features and those agents will be able to negotiate and transact. We will have tools to help Ethereum developers move apps and technology across, and “agent construction sets” that make it very easy to construct and deploy agents with different features. In short, it’s the first safe, secure, stable version of Fetch.ai’s technology that you can do business on. And yes, there are many use cases. A good few of these you will be seeing before the launch of the mainnet.

How did your team deal with the technical challenges during the initial launch of the project? What milestones can we look forward to and what challenges may you face in the future?
The technical challenges were considerable, and we deliberately operated in stealth mode until we had dealt with the key innovations. When we originally published the white paper, a video and announced Fetch.ai to the world, we had already proved the core smart ledger innovations would work and that they would provide the performance we would need. We also had invested time and money into the AI/ML side of things so that we knew that key AI features such as semantic views, reputation and trust information, etc., would behave in the way we expected. Key future milestones include the mainnet at the back end of this year and a ton of very cool decentralized AI applications coming from now and throughout next year.

After the AMA finished, Humayun selected what he considered to be the best five questions. The individuals who submitted these questions will each receive 2,000 FET tokens, taking the total amount of tokens airdropped to AMA participants to 130,000 FET. The winners of Segment 3 are:


Thanks to everyone who participated in the AMA. If you weren’t lucky enough to win tokens this time, better luck next time. We hope you learned a lot during the AMA about Fetch.ai’s decentralized vision for the future.

If you would like to learn more about how you can build on the Fetch.ai network, we strongly encourage developers to join our developer Slack channel. If you would like to keep up to date with all of our latest news, please join our Telegram channel and follow us on Twitter.