1

In Conversation with Amir Haleem, CEO, Helium

 2 years ago
source link: https://mattturck.com/helium2022/
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.

In Conversation with Amir Haleem, CEO, Helium

In a crypto industry that outsiders often like to criticize for its supposed lack of clear use cases, Helium stands out. The New York Times recently recognized this reality in a recent article, saying “Maybe There’s a Use for Crypto After All.”

Helium is a decentralized wireless network, powered by cryptocurrency. Mostly focused on powering “internet of things” devices for now, it’s been rapidly evolving towards 5G.

The Helium network has experienced remarkable success over the last few years – it’s built a global network of almost 670,000 hotspots deployed around the world (see discussion to understand more about hotspots).

As most “overnight successes”, however, Helium has been many years in the making. I’ve had the honor of being part of (almost) the whole journey, as I wrote on behalf of FirstMark the first institutional check into the company back in 2013 (what would be known today as a pre-seed), and reinvested a number of times since.

Along the way, we’ve had Helium speak at our FirstMark events several times, which is a fun reminder of the journey: then CTO Sean Carey in 2014 (here), and CEO Amir Haleem in 2017 (here) and 2018 (here).

So it was great to welcome Amir back once again to chat about the latest.

Below is the video and full transcript.

(Crypto Driven is a team effort – many thanks to my FirstMark colleagues Jack Cohen, Karissa Domondon and Diego Guttierez)

VIDEO:

TRANSCRIPT [lightly edited for clarity and brevity]

[Matt Turck] What is Helium?

[Amir Haleem] Helium is an economic system for building wireless networks. Today, the first instantiation of that idea is an IoT network. When we started Helium a long time ago, back when I think you first met us, Matt, in 2013, we believed that IoT was the worst solved problem in telecom. There were no solutions for connecting devices to the internet that were small and battery powered and needed to be cheap.

And so, our focus for the longest time has been on solving that problem. How do you build a big ubiquitous global scale IoT network? We took lots of different attempts to try and solve that problem as you well know. It was not all up into the right. It was a very twisty road that often went left and downwards and went all over the place.

With crypto, we had the light bulb moment somewhere in 2016 or 2017, where we realized that you could empower people, whether they were individuals or small businesses or even large businesses to participate in the telecom industry and use crypto as a way to organize that. Helium is really that. It’s the economic system that allows individuals to participate in a global telecom network.

The first instantiation is IoT, but it certainly is not the last and we’ve moved already into 5G and cellular and there will be all sorts of other ventures in that direction as well, whether they are WiFi or fixed wireless or whatever other future version, but it’s really that. It’s like enabling people to coordinate at a global scale to build wireless networks.

Helium has a supply side and a demand side. How does it all work?

With IoT in particular, you’re starting from a blank slate. Even though IoT or the internet of things has been something that has been talked about for 10 or 15 years, there really hasn’t been a good solution for how do I build something that connects to the internet this way? And so, this is one of those things that you’re not supposed to do at startups. You’re not supposed to do the “build it and they will come” approach, but crypto allows us to take that approach. And so we did. The way we chose to solve it was to say like, “Okay, we have to build a supply side out first.”

The supply side in our case is network coverage. Today, when you have a cell phone, you walk outside or you don’t even need to walk outside, inside, you have coverage everywhere with cell phones. We wanted to build the equivalent of that for IoT devices. Sensors and small things. The solution there for us was to create a device called a hotspot. The hotspot in our case is a combination of a crypto miner and a wireless access point for IoT devices.

It uses a wireless network protocol called LoRaWAN designed for very low power things over very long ranges, very low data rates, so you’re not streaming video or anything like that. You’re just sending relatively simple messages like GPS location or something like that. The thing that I think was special about what we did with Helium is that we made it very simple to do two things that were otherwise quite complicated to do for consumers. Like mining crypto, not that straightforward.

If you’re in a CLI or you’re buying an Antminer, which is like a shoebox full of wires… none of that UI is designed for consumers. The same goes on the other side, trying to set up an IoT access point, also not designed for consumers. These are designed for telecom companies and the UI is also atrocious, and you need a 48-volt power supply. And you’re in a CLI somewhere. We managed to combine two things that were not consumer friendly into a thing that was quite consumer friendly.

And I think that has caused it to succeed. So, really, all you have to do as a consumer is just get a hotspot, plug it in, use a phone app to set it up, which is a similar experience to any consumer electronic device you might buy. And that’s it. That’s your participation process. I think making it that simple and that easy has caused the network to explode. It’s like 600,000 hotspots. Now it’s like three and a half million that are back ordered. It’s in thousands of cities, hundreds of countries. It’s really taken off in this way, I think, because we’ve enabled pretty much anyone to participate in this world.

What about the demand side of the network? What are the emerging use cases that you’ve seen?

The demand side has always been the slow part of the equation unlike in software where it is relatively straightforward or certainly not as time consuming and from an overall start to finish point of view to build applications. Hardware is much more arduous. You’ve got to build a prototype. You’ve got to figure out how to manufacture it. Every iteration of the hardware takes weeks or months.

Everything about that process is slow. And so the bet with Helium was always that if we built the network, that the use cases would appear, basically, which I know sounds kind of insane from a normal business point of view. But our experience over the years had led us to believe that use cases existed and there were lots of them. There was just no way to deploy them. It’s almost like if you had the idea for an iPhone app, but the only phone you had was like a Nokia flip phone. There was no way for you to make it occur.

And so, we’re starting to see that now appear, the demand is coming and some of the use cases are super interesting and we’ve announced some of them and we haven’t announced some of them. But whether it’s things like delivering packages via drones that need coordination or whether it’s like connected mouse traps or GPS tracking devices, which I think was always the most obvious application for things like logistics and pharmaceuticals and everything else, wildfire sensors. There’s a very, very broad range.

We’re starting to see them appear now whether they are from bigger corporations that already had the idea or whether they’re from startups or individuals and entrepreneurs that are coming up with new ideas. But to me, that’s the most exciting part, is to actually see the application start to form and take advantage of the coverage network that exists.

You built your own blockchain. Do you want to talk about why and how, and then how it actually works, including proof of coverage, proof of location, the whole consensus mechanism at a more technical level?

The time that we conceived the crypto version of Helium, which was ’16, ’17 timeframe, Ethereum was really the only way that you would feasibly build something like Helium, if you were going to do it as a layer 2. We already thought that that was going to be a problem. Gas prices were already quite high. We knew that the number of transactions was going to be significant in a network like Helium.

And so, we weren’t sure how to proceed, whether you would have some state channel implementation or whether you would try and roll up as many transactions as you could, the way that a lot of the Ethereum layer 2s work today. The other thing people were doing was forking Bitcoin. You could take the Bitcoin code and fork it and run that way. Everything that we wanted to do didn’t really fit into either of those.

I’m glad we made the decision not to run on top of Ethereum because I think it would’ve been really, really challenging to make that work and the cost of it would’ve been probably impossible. It felt like by necessity, we were driven to build something. Certain members of the core team had distributed systems backgrounds. They’d worked on things like the iMessage backend and we felt like we had enough people with enough experience to actually do it.

And so, the Helium blockchain is a layer 1, it’s a proof of stake network, where there are validators who are responsible for transaction validation and block production. It uses a consensus protocol called HoneyBadgerBFT. BFT is the Byzantine Fault Tolerance world of consensus protocols and the HoneyBadger version of it is an asynchronous version that doesn’t really care about latency. When we thought that was important at the time. Unclear if it’s still super important as the networks have evolved.

The validator group, there’s about 3,500 validators, about a third of the entire HNT supply is staked. It’s proof of stake blockchain, and you have to stake 10,000 HNT to become a validator. There’s about a billion dollars worth of HNT staked in all these validators today. And the rest of the network uses something called proof of coverage to figure out who to reward. One part of the system is creating blocks and validating transactions. And that looks just like any other blockchain for the most part.

But the other in our case was, what are we trying to do with the Helium network? And really, the thing that was special for us was rewarding people for building the network coverage before there was usage for it. We devised this scheme called Proof of Coverage, which is basically hotspots transmitting encrypted packets over the airwaves to each other. And the network would try and reward people who created the most coverage.

So if you could prove that you could hear hundreds of other hotspots at very long ranges, then you would earn more than someone who had it in a living room on the floor that couldn’t hear anyone. Proof of Coverage is designed that way, it’s quite detailed in terms of how it works and quite complicated, but the gist of it is the longer range, the better your coverage area, the more HNT you earn.

Let’s talk about the tokens and how it all works in the tokenomics. There are HNTs, there are HSTs, there’s the concept of Data Credit. How does that all work?

The main token is HNT. That’s the native, the token that hotspot owners earn. It’s what validators earn, it’s what’s used for staking. The purpose of HNT is to be able to create Data Credits, which is the second token in the Helium ecosystem. And Data Credits are, you can think of them as a stable coin inside of Helium. And they are similar to something like a cell phone minute or an airline mile or something like that. It’s utilitarian in purpose. It allows you to transfer 24 bytes of data over the Helium network wirelessly. And that’s all you can do with it.

You can’t trade it. There’s no speculative value to it. You literally can’t even transfer them. The only thing you can do with them is destroy them to pay for traffic. We like that system. It was actually influenced heavily by Kyle and the Multicoin team when we first met. And we like that system a lot because it gives predictability to end users, but the value of Data Credits is always the same. And so if you’re a user of the network building sensors, you know exactly how much it’s going to cost to use the network at all times, regardless what the value of HNT is, which is fluctuating around.

And so you can only create Data Credits by destroying HNT, that’s the… They call this the Burn and Mint equilibrium. You are destroying HNT in order to create Data Credits. And the ratio changes based on the value of HNT. Data credits are fixed in value and HNT is fluctuating in value. That system works really, really well because as I said, it gives very predictable outcomes for users of the network and no one has to second guess what the price of using the network is, which I think is a problem in systems like Filecoin and others that have a single token.

The third token is not that relevant for users other than it being an interesting founders reward system. We have a third token called HST, and we modeled this a little bit after Zcash. Zcash was the only system that we knew of that was rewarding the founding team and its investors over time, rather than in a pre-mine. Most blockchain networks will create all of the tokens up front and they will allocate a percentage of them to the founding team, its investors and that’s the way it works.

And we never liked that mechanic that much because it always felt incorrect. It was slightly unfair to just give the founding team all of these tokens up front and let us sell them whenever we wanted. It never quite felt right. And so we used the system, which we borrowed from Zcash, where HST holders would earn a percentage of mining rewards as they were created over time. We thought that was more fair. The difference between Zcash and Helium is that HST is tradeable and Zcash founder rewards are not tradeable, but that’s basically the three sets of tokens, HST, HNT and Data Credits.

To start weaving in some questions from the groups, a question from Nitesh about effectively how you see the incentives evolving. “Incentives were the key to scale the Helium network, but with growing device numbers and lower rewards, how do you conceive of incentives going down when we need to keep scaling the network?”

It’s a good question. I think part of what happens in crypto is that people become desensitized to normal rewards or normal ROIs. And that’s a little bit of a messaging challenge. Early hotspot hosts by today’s HNT value made an absurd amount of money, hundreds of thousands of dollars, even millions of dollars in some cases for their earliest miners. I think that distorted what acceptable ROIs and acceptable rewards look like. And so, as far as I understand, HNT mining is still one of the most high ROI miners that you can participate in, uses no power, has no real running cost.

I think to me, it feels like we’re still far away from that problem, but there is a sort of expectation issue. There’s a gap between 50% APR being good and people being upset. And I’m not sure that we’ve bridged that perfectly or that the community as a whole has been able to bridge that properly. But there’s a few things going on that will help over time. Like hotspots are going to get cheaper over time. I think this is a really bad period of time for anything to do with hardware and manufacturing over this COVID period.

There’s an upgrade to the network called HIP 55, which will cause the hardware requirements of hotspots to go down significantly, so it’ll be much, much cheaper to manufacture and produce hotspots. All of that changes the ROI calculation or the incentive calculation there. As hotspots get cheaper, you need to earn less and you need to be less motivated and incentivized to participate.

And so, I think over time, things will settle down and as usage of the network picks up, I get less and less worried about this, but I do think there is going to be a constituent of earlier users that have made all their money and now start to cycle out of the system and that’s perfectly normal and perfectly expected.

While we are at it, talking about the tokens, do you want to talk about the decentralization aspect of this? Or for example, you mentioned the hardware and the requirement for the hotspots. I think it’s important to know that Helium is not in the hardware business as you decentralize that to the community. That’s one aspect and maybe talk about how the network is structured from a foundation perspective and other things.

Yes. There’s something like 65 different hotspot manufacturers now. We built them originally not because we wanted to be a hardware manufacturer or we wanted to make money. And I think a lot of people looking at Helium mistakenly think that Helium, the company earns anything from hotspot sales. People often say, “There’s a licensing fee. Helium Inc. doesn’t care. They’re just interested in selling hotspots.” We actually make nothing from any hotspot sales. There’s no licensing, there’s no nothing.

And so for us building the first hotspots was important because it allowed us to create the user experience that we thought mattered the most. Everything from that has been open sourced and handed off to third parties. The hotspot ecosystem is as decentralized as it could be. We actually, as a company have no interaction with the vendors other than to assist with technical questions and technical problems over time. Other parts of the ecosystem are similarly decentralized, like up until very, very recently the company, for example, didn’t run any validators.

There are now 3,500 validators, 3,300 of them are nothing to do with us. They’re members of the community and other people that have bought HNT or participated in the ecosystem. It’s heading in the right direction. I think what’s hard about the decentralization story with blockchains is really the core engineering that needs to happen. And you see this with Bitcoin and others where there’s always relatively a small group that are skilled enough to participate in the engineering.

I would love to see that grow in a more decentralized way and have more input from the community, from an engineering point of view, but as far as hotspot and validators and users have gone, I’m really happy with how distributed that is and how little part the company really plays in any of that anymore.

What about governance? And you mentioned HIP, what is it? How does the community impact the development of Helium?

There’s a process that we borrowed from Bitcoin. Bitcoin has a thing called Bitcoin Improvement Proposals, which are BIPs. We have Helium Improvement Proposals, which are HIPs, and anyone can submit a HIP. Nearly all of the changes on the Helium network have come from proposals from the community and not from the company itself. There’s a few exceptions, but for the most part, they are community driven. And so, someone from the community writes the proposal, it gets voted on by the community. And it’s a HNT weighted vote.

Recently, I think the network is to the point where these HIPs now have to really come with code. I think we were getting to the point where people were writing extravagant and often very, very good proposals for how to improve the network. But there was no one to actually like do the engineering work. And so, I think the network is evolving now where the expectation now is if you want to submit an improvement proposal it needs to come with the code that comes with it, and that’s similar to the way Bitcoin and Ethereum and other networks do it. It’s like, here’s the change I want to make and then there’s a debate and discussion and a very active process and then people end up voting on it and either it’s accepted or rejected.

Question from Chris in the chat, “How did you navigate whether or not HNT and HST was considered a security by the SEC?”

The trap question! [Laughs]

Yeah. I’m picking the hard questions just to be very clear. [Laughs]

Candidly, I wish there was clearer guidance on how to proceed in some of these, but I think we were very careful about the token design here and we never sold any HNT, there was no pre-mine of HNT. I think we did as much as could be done to ensure that we were on the right side of the regulations there.

Question from Keith. “When can we get delegated staking please, really, really, really please?”

The question here just for context is right now, you have to stake 10,000 HNT to participate as a validator. And that today is a quarter of a million dollars and at times it was even more than that. There are third party vendors in the ecosystem where you can stake much smaller amounts. You can stake a hundred HNT for example, but the downside is that you have to trust the vendor. I would love to see partials staking on chain and I think it’s just one of these things where there are more pressing concerns, I think ahead of us. So far the ecosystem of third party vendors seems to have been trustworthy and good. I would recommend that for the time being.

Question from David, “Will you please speak to the recent vote to undergo a thousand X redenomination?”

I’m not sure there’s really… That did not come from us. A community member proposed splitting HNT or not splitting, redenominating HNT at a thousand to one ratios. So if you had one HNT in your wallet today, you would have a thousand after the change was made. I think it’s interesting. I do think there is, what people call unit bias in the crypto world, where I’ve had hotspot miners directly tell me that I would rather earn a thousand HNT that was worth nothing than one HNT that was worth a hundred dollars.

And so, there’s something strange that goes on mentally, where I think people like large amounts and they can conceive of them being highly valuable in the future. To me, that’s where that idea came from, was someone proposed that redenominating HNT it passed, and I think it’s going to happen later in the year. It just, again, hasn’t been a priority for us to implement in the near term, but as far as I know, it’s still coming later.

There’s various questions around the general theme of where is this all going. Question from Samuel, for example, “Realistically speaking, where do you envision the Helium ecosystem will be in two years, and then five years? When do you envision that Helium’s 5G flywheel would be implemented?” There was another question on the 5G as well. Related question for Keith. “Do we need to cover every inch of the earth with LoRa network, or is it okay to cover most of it from a fuzzy perspective?”

Question from Andrew, “When the network is fully built out in most developing emerging countries, what is the estimate for the total amount of Helium hotspots? Do you need to cover everything? How many hotpots is that going to be?” And then as a separate question, and one or two others ask as well, “What is the timing for 5G?”

I’ll answer the 5G stuff first. The hardware for it exists and it works and it functions, and you can buy a 5G hotspot today. And I think GigSky is the first MVNO that is using the Helium network. You can be a GigSky subscriber and participate in the Helium network already. I think really, the question is when do I start earning proof of coverage rewards for 5G, or when do I start earning rewards in general for 5G?

And I think I would be hopeful that that’s this year. We’re working with the FreedomFi team on a regular basis thinking through that, and there’s various proposals in the community. Again, I think there’s HIP 37 and 51 and 52 and 53. Lots of different suggestions here in the community about how to implement rewards and like how to enable mining in the 5G universe. I’m optimistic that the network is on the right path there.

How does backhaul work with 5G as part of the department strategy?

Today, backhaul is the same as it is for the IoT network. We expect users to bring their own backhaul whatever that may be. I saw one of the questions in here is about Starlink, for example. Totally valid backhaul to use. The network doesn’t care how you’re connected. It does care that the connection is good and reliable and stable, but beyond that, do whatever you want there. And I think that’s an important thing going forward. And as we see diversity of options for backhaul, I think that’s going to become quite exciting.

Already it’s feasible through something like Starlink to create cell networks literally in the middle of nowhere. And I think that’s amazing, and I think we’re going to see evolutions of that and we have various conversations going on with potential partners about how that might work in the future. In terms of some of the other questions about LoRa and where will Helium be in two years, five years. I think it’s entirely feasible in the next two to three years that Helium is the largest cellular network in the United States.

And I’ve said this before, and I think lots of people in the industry think I’m crazy, but it’s entirely possible that that’s true. If you look at the size of something like Verizon or T-Mobile, their cell networks are something like a hundred thousand towers in the entire of the US, and there are already 600,000 hotspots with another three million coming. So if you see anything like that success in the cellular in the 5G world, then it’s entirely feasible that Helium gets to a size where it is the dominant cellular network in the US in terms of coverage.

That will be very disruptive from an economic point of view, I think there will be all sorts of other carriers and MVNOs that get built just purely on top of Helium. I’m beyond excited to watch that unfold and see if it’s actually possible to do this. I think having CBRS as an option is such a big deal because that’s really been the moat for the longest time, is that you couldn’t buy spectrum as a small size company, there was no way to buy $50 billion worth of spectrum, but now you don’t have to in order to start building.

I’m excited about cellular, especially, I think the demand side is obviously solved relative to IoT. We know what the predominant use case is for cellular networks where that’s not necessarily true in IoT yet. And now the problem is just building the network out and making sure that the incentive is correct in the same way it was for the IoT network and I think we’re on the right path for both of those things.

Question from Benjamin, “As Helium grows and launches a 5G mobile service, what makes you think that internet service providers will not cut off Helium devices from their network?”

I think it’s probably the most common question that I get asked. I continue to just view that as not a thing that is likely to happen. To me, it doesn’t feel any different from using my home internet connection right now to conduct Helium Inc. business as I’m talking to you. I don’t think that violates any terms of service that I’m aware of, but I’m conducting business and I’m earning money. To me, they are one and the same. In general, people need to look at their own terms of service and make sure that they’re not doing anything they shouldn’t be doing, but on the whole, that’s not an issue that really ever concerned me.

Question from Spencer, “How does the rest of the telecom industry view Helium?”

God, I’m not in the rest of the telecom industry. I would imagine that mostly they’re intrigued at this point. We have good relationships with some of the carriers, we announced with DISH recently and we work with other carriers on the LoRa side. So certainly everyone is I think some combination of impressed and intrigued. I think there are a whole group of people that still think Helium is a Ponzi scheme and crypto is a joke and whatever, and I look forward to seeing their reaction over the next few years.

Maybe to wrap up some of these questions and just in general for anyone listening. What are the different ways one could get involved at Helium? There’s the mining, there’s buying the tokens. There is a question around if you are an entrepreneur and wanted to build something on top of the network, what would be interesting from your perspective? From the very mundane, easy to actually building a company on top of the network, what are the different ways one can work with Helium?

It’s pretty broad. If you go to helium.com/ecosystem, I think that probably gives the best view of the different entities that participate. But as you said, it’s typically either hosting a hotspot, buying HNT, you can operate a validator. Those things I think are always going to be high value and useful to the network. But what I’m personally the most interested in seeing nowadays is applications and use cases that people build. So, if you go to helium.com/use, there’s a helpful guide there on how to get started.

There’s lots and lots of cheap development kits available. I don’t know. In terms of use cases, I get asked that question a lot too, and I don’t know. I’m constantly shocked at some of the things that I’ve seen and I think that’s the best place to be. And again, it reminds me a little bit of the iPhone. Could never have predicted some of the applications that now exist on the iPhone. I can’t predict the Helium applications either other than thinking that they’re awesome when I see them.

That’s, to me, the thing that I’m always the most interested in is what people build and seeing makers and entrepreneurs coming up with new products and ideas that use Helium is to me, one of the most exciting things about the network.

Very cool. Well, it’s been an incredible journey, and in some ways it feels like Helium is just getting started. And I know for a fact, we’re all incredibly excited about the potential and what’s in store for the network. So congratulations on everything you’ve built has been incredible to watch and so excited for the future. And thank you very much for coming to speak at this event. Really enjoyed it. Thank you.

Yeah. Thank you. And thanks for having me.


About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK