Introduction
Hey 👋, I’m h5law. I’m 23, I am on a one year foundation degree in Computer Science at university - a full 3 year course seemed pointless to me in today’s times. I will probably say this a few times but in Web3 its not about who you are, but what you can deliver. If you can learn something yourself and get to a point where you are able to ship feature after feature without any degree, I see that as a win. This last year has been a wild ride for me, and as a “memoir” of sorts I thought I’d write down some of the highlights.
So in addition to being a student I am a protocol developer at Pocket Network working on building out the new version of the protocol - referred to as “Shannon”. We are building a permissionless, decentralised RPC network that incentivise API services in general. The main focus at the moment is for Web3 projects, Blockchains and dApp developers but the longer term vision is for Pocket to be the base layer for RPC access to any and all API services (LLMs, and whatever else you can run we could service your requests).
This article is about my journey over the last year, after submitting my first PR a lot has changed. I’ve learned so much; grown as a developer; made/contributed to some really cool stuff, in my opinion at least; and been challenged in so many ways. I hope to capture some of these in this article
Context
For those who don’t know, in the past year Pocket has transitioned from building out its own L1 to a rollup utilising Rollkit and Celestia for our data-availability layer.
When I started I was building lots of new things, below is a list of things that stood out as of the time of writing:
- Crypto and Key primitives.
- Keybases.
- Hierarchical Key Derivation Systems.
- See: SLIP-0010 for more information on this specifically.
- The CLI and RPC to interact with the L1 itself.
- A non-cosmos dependent implementation of ibc-go
- This was very much still a WIP
But it was when I began to work on the SMT that I found a true passion of mine. We will dive into this in a lot more detail later, but it was in the old Morse (v1) repo this all began.
After much deliberation, many late night calls, debates and most importantly
research; we decided as a team to move away from building out our own L1 and
focus on our core offerings. Decentralised RPC. Olshansky wrote a
great blog post
detailing everything so I won’t go into it all now. But as a result it meant
starting over with: a new stack, a new codebase and unfortunately a lot of the
work I had previously done was scrapped in favour of the multitude of things
that come for free out of the box with using cosmos-sdk (rollups
using rollkit on celestia can still use the sdk as if they were building a
regular app-chain) and benefit from all the work done in the cosmos ecosystem.
So lets get into this last year: the good, the bad and the ugly.
The Good
I have honestly learned so much, its unbelievable how many challenges you face when building something completely new, from the ground up. It’s true you shouldn’t re-invent the wheel but what if the wheel never existed in the first place? We have had to make so many cool things that honestly make me so proud to work with such a great team and build such an amazing and innovative protocol.
From the guidance of Olshansky and the other amazing members of the protocol team Bryan, Redouane, Dmitry - all who have a lot more industry experience than I do; I have managed to build my intuition on problem solving, which is the hardest thing (in my opinion) when it comes to programming. Writing code is easy; debugging is hard; testing is a ball-ache; but the way you approach a problem, especially one you haven’t encountered before is by far the most challenging aspect of development.
As part of building out the Shannon upgrade to the protocol, I am still involved heavily in the cryptographic side of things. Integrating ring signatures for our App to Gateway delegation process and request signing. I’ve been able to contribute back to the library we use, which I find so cool.
So lets break down some of the highlights of this year. Of course there will be lots of exciting things I am going to miss out, otherwise this blog would be incredibly long.
SMT
Over this past year I have learned a lot. I have become somewhat of a specialist in the field of trees/tries 🌴 with the SMT repo that I have taken over, since celestia archived their repo which we were using previously.
Since taking over I have merged in numerous features:
- Lazy Loading (developed by celestia but never finished)
- A Merkle Sum Trie
- A possibly world first implementation of a (Sparse) Merkle Sum Trie implementation, based on the plasma specification
- The
ClosestProofdeterministic commit and reveal proof mechanism- This is a novel proof mechanism based off of the Relay Mining Paper
Documentation for all of this can be found here.
We had our first external PR and as such discovered IOTA Ledger are using our SMT! It really felt like my baby 👨🍼 was taking his first steps.
I’ve got so many plans for this repo and its future, that I will save for another post but all I should say is:
Expect big things 😉
Compute Units
If you are familiar with RPC providers in general you would know providers such
as Lava, Ankr and Alchemy utilise Compute Units 💾 these allow for each RPC call
to be paid for proportional to the cost required to execute it. This is beneficial
as it rewards suppliers for providing more demanding services and thus gives
more access to those services.
I am leading the work on integrating Compute Units into Shannon, this is one
of the main reasons I made the Sparse Merkle Sum Trie (SMST) as we can attach a
weight or sum to each relay & response pair a supplier inserts into their
trie and easily get an accounting for how much work they did.
It is super exciting to be working on getting this feature into the protocol, with out-of-the-box ideas instead of the typical verbose approach one would immediately think of. Again big shout out to Olshansky and Ramiro from PoktScan for helping me flesh out these ideas before we begin the implementation.
Gateways And Rings
I implemented the on-chain gateway actor and in conjunction with Redouane built
our MVP off-chain appgateserver (which is essentially a gateway like Portal
or NodiesDLB). We are currently working on building out an SDK to allow for
Gateways and applications to interact with the chain. All of this is under
active development but seeing some of the issues with Morse (v1) being
eliminated is actually amazing and I am really honoured to be a part of it.
My main feature of the interactions between Applications and Gateways was
(un)delegation. This allows for an application to delegate trust to a gateway
such that the gateway can sign incoming requests from the application on its
behalf. All of this works through the usage of ring signatures 💍✍🏻
and the RingCache which handles the creation and, obviously,
caching of rings on behalf of anyone who requires them, (applications and
gateways for signing request and suppliers for verifying requests).
Cryptography has definitely stood out to me as being another area of interest, and I hope that soon™️ , I will be on the same level of knowledge and intuition with cryptographic primitives as I am with trees. But there are lots of abstract algebra textbooks 📚 to read before I am fluent here.
The Bad
Naturally as part of pivoting: a lot of work I was very proud of; ongoing work
related to a native go implementation of IBC (without the cosmos-sdk
dependency) and other parts of the old repo I was invested in suddenly vanished.
It’s never nice to see the things you’ve worked on and eagerly waited for people to use not see the light of day. But that’s life and we move on. As the section above details there is still so much good that this doesn’t really matter. What we are doing now is the right thing and I stand behind that with my whole self.
Personally, I find it hard to find any downsides to my work here at Pocket. Going from a community contributor to working full time on the protocol, I live and breathe POKT. I love this project, I have ideas 💡 ranging from pre-testnet, to testnet and after mainnet. And I want to see all these implemented. It’s been great to have my say and shape the protocol - or at least influence its shape.
That being said I find coming from a different background to the others on the team to be a real challenge. I am young, have less “experience” (but this doesn’t matter in Web3 - if you can deliver it doesn’t matter who you are), and this shows. Often its an opportunity for me to learn, other times its an opportunity for me to teach them something. We are building out the protocol rapidly and everyone is committing their all. Working weekends, late nights etc. But occasionally there are miscommunications and this is expected working in a team. Maybe its because I haven’t worked in the “traditional” corporate setting, but it seems my style of work doesn’t fit well with others. I think the new generation hustles, grinds and works differently from the generations that came before us. The rest of the team work hard; they grind; they hustle; just differently.
This could be the single downside I have had working full time at Pocket. Going from a community contributor to a full time core protocol dev, it has become a job, there are more rules; people to please; feelings to take into account. In my world all that matters is:
Can you deliver? 📬
The Ugly
As I said previously, I am at university 🎓 full time as well as doing all this work with Pocket. Surprisingly this isn’t the easiest thing to do.
Doing both university full time (with all the: exams, coursework, lectures, practicals etc.), on top of the full-time work I do at Pocket, has been a lot. It has definitely been ugly 😖. But I’ve achieved time management skills like no one would believe. I surprise myself sometimes with how I manage it all. It’s far from easy but its most definitely the ugliest part of the POKT+UNI combination.
I think I’ve come close to burning out a couple times, I’m definitely someone who gives 100% to whatever I do and this level of intensity is tough. But I love it. The title of this section may be a little misleading in that regard. It’s tough, it’s hard, but there’s nothing I’d rather be doing than what I am doing right now.
Summary
In summary, I live and breath POKT. I love working on and building out such an innovative protocol. Being part of something that I know will be huge and used by so many people, devs, projects, companies, etc. excites me. It drives me and motivates me to continue working on this amazing project.
The interests I have developed working on POKT: Trees 🌴, Databases 💾,
Cryptography 🔑 will stay with me and I will continue working on building the
best solutions not only for our use case but for the entire industry. I truly
believe that we will soon enter the PoktVerse 🪐 phase:
- We underpin so many projects as their infrastructure and data provider
- Our innovative technology is used across the entire industry
- We moon? 🌝
In Web3 (and Crypto in general) it doesn’t matter who you are. All that matters is what you can do. Your past isn’t important, neither where you have come from. I think this is the best part of the industry. And I’m really looking forward to what comes next, so pay attention 😉.
