← Writing

2020 Vision

October 1, 2020

Quick Update: Summer 2020, Crypto 1.0

Back in school, taking 18 units, daydreaming about MEV.

This was my most productive summer so far. I interned with the consumer onboarding team at Coinbase, hacking on react native and typescript. I picked up some cool hax and tricks, met a bunch of smart folks, and am back in Defi.

Back in 2019, I went over some things I'm interested within crypto:

Market share amongst emerging smart contract platforms: Turns out my gut won this one. Ethereum is by far, the most stable and use-out-of-the-box kind of solution. I see what NEAR, Solana and others are trying to do but don't see them succeeding anytime soon (I'm also pretty biased towards eth, don't see the need to redo an entire chain just so a few peeps can capture value)

Crypto/Blockchain as a utility: We're seeing some growth here, but real sluggish. Outside of defi, crypto provides a lot of opportunities where you can verify ownership etc.

Credibility/Establishing Trust: This is still a valid problem. In 2020, it is not possible to verify/trust/cross-check an ethereum address. 3Box sounded like a good solution, but hasn't grown.

I'm interested in seeing how you can combine good-faith actions on-chain and track them. This could open up a new design space where access to dapps is based on an the good-faith activity of an address on-chain.

So what is a good faith action?
  1. Liquidation of accounts on money markets: You're helping a protocol stay solvent and function properly
  2. Providing liquidity: This has been done (with COMP, and the rest) and is a huge milestone in taking crypto forward.

Still thinking about this one, but liquidations seem to be a great starting point. HMU if you think of anything.

Now for the most interesting part -- DeFi

From telegram to a niche: DeFi 1.1

DeFi seemed largely useless, with okay-ish but better than traditional finance interest rates. Few understood the value prop, the ones who bet on code commits vs. hype cycles won this round.

At its core, DeFi shows us what trustless interactions can bring to the table. Secondly, the DeFi gold rush of 2020 is fairly distributed (not dominant by valley based companies) and largely focuses on a new tactic: Exit to Community

We all know that theres a certain big fund pushing crypto as an industry forward, and there are lots of alternative funds beyond this. A16z now holds the opinion that successful, and progressive decentralization is the correct way to build web3 protocols, which I believe is going to be the narrative for the forseeable few years.

Code: What I've learnt 1.2

Lots on this side. To begin with, I tried a lot of cool techs: GraphQL, Typescript, Python3 + asyncio, and a little bit of go. Still feel the most comfortable with js, typescript can be a timesink at times, and python3 is clean for discrete one-off problems.

I'm beginning to move into deep eth (yes, dark forest like) development and have been playing with seth lately. I find the mempool fascinating. This fascination began when I discovered dydx and on-chain liquidations during my internship last summer. I started tracking data and found the number of liquidators exploding, and found this to be a race-condition (i.e: first come first serve) and an interesting problem worth jamming on.

This, coupled with the flash-loan exploits in February this year, and the unfortunate sequence of events during Black Thursday (March) have rekindled my interest in the inner workings of eth.

At its core, ethereum is an adversarial environment, and transactions are getting frontrun (and backrun) left and right. I'd like to dig deeper and find out whats goin on in the mempool and explore it.

Current toolkit:
  • Lerna: I use monorepos for almost everything now
  • Tmux + Nvim: A brave attempt to ditch IDE's
  • GraphQL: I try to use this to stitch apps together, and also because I end up using thegraph a lot to quickly pull on-chain data. In general, I’d like to move away from REST and use graphQL exclusively.
  • Python: Few of my projects required me learning asyncio, and I now see how I can use this
  • Devops stuff: I built a system in kube this summer. Amazing experience and devex, will be digging deeper into docker so I can setup a fullblown EC2 instance for personal uses.
Crypto from here 1.3

I think we see new protocols, more complex financial instruments, and tremendous amounts of capital slowly pour into DeFi. Of course, this also means we see more of:

  • Scam bots duplicating and deploying Uniswap tokens and markets (noone can stop this given uniswap can't do much beyond providing lists)
  • Rugpulls, spectacular fuckups and smart contract bugs
  • Desperate calls out for attention (hello! eth killers, battle of the maxis)
  • Amazing irrationality (crypto is by default)\
Protocol Wars

Not sure if this is the right title, but the forkable, easy to clone nature of DeFi protocols (most of them are open source by default) opens up a new attack vector. Building a moat is going to be the #1 thing to optimize for, else protocols are going to get rekt ('Chef Nomi' dropped some wisdom earlier).

On a serious note, this is a fascinating area thats probably underexplored. Mechanism design and economic modeling really kicks in here. The goal is to answer a simple question: How do I ensure user 'X' sticks to my protocol vs. my competitor

The answer: not so simple, involves incentives, and tight security measures. We're quite literally scratching the surface area of whats possible.

If compound finance still has 800million USD locked, is it because of the brand, the fact that they've pioneered on-chain money markets, or because they've never been hacked and have a strong community? The answers probably a mix of all of them.

After several discussions over the summer, and listening to some cool podcasts, I've come to the conclusion that populism is pretty visible and a real thing within crypto. Theres the SNX Spartans, LINK Marines, and what not. The community sticks through rough patches (bear markets, protocol operating hiccups etc)

On the brighter and better side, we're going to see crypto grow by a tiny margin, and take us closer to what this niche aims to achieve as a whole. No, this doesn't mean 'decentralize everything', it means that builders are going to iterate and present to a larger audience, value accrues to protocols actually generating fees (uniswap > many chains combined right now) and we'll see a renewed importance towards anything on-chain

My current area of exploration is the mempool, guess I'll see you in the dark forest eh?

Defi as a design space, crypto-natives, COVID 1.4

This year no doubt has been a rollercoaster. While not diving into the other details, we've quickly embraced remote-culture, defaulting to zoom, google meet and the likes.

COVID to me is a forceful experiment asking people to spend more time online. I think about the implications and changes in online behaviour. No, I'm not talking about tiktok (although interesting from a gen-z perspective). I think this opens up new avenues: produ tivity focused apps (e.g: superhuman, linear for anything)

Online Games & Crypto 1.5

(Sorry, we're not discussing NFTs. I think they're cool.. but not today)
I spent a good part of being a kid playing games like Runescape, farming stuff ingame, flipping them on this 'Grand Exchange' which is a kind of a limit order book for any item in the game, trading accounts, and learning how to type fast ingame.

Other games like CS:GO have the same characteristics. People come for the game or the hype, and they stay for the uniqueness, drops, merch and a desire to flex on others. This is generally how humans behave. They're addicted to exclusivity in a way.

CS:GO had skins (paint on top of whatever gun ya wielding), and a variable (float) attached to each skin. The lower the float, the more rare and cleaner your skin looks in-game.

There were alternative markets, where you deposit your skins to a bot, which then lists it on a private site, frequented by collectors and traders. A subreddit, and ingame servers helped you meet, negotiate, and trade.

To me, this sounds like a hacky fix to a problem (you're stitching together reddit, self hosted versions of CS:GO, and unofficial sites to create a trading environment), which could probably be fixed by crypto. Worst case: learn from the activity on here and adapt it into crypto.

I'm going to try editing all my drafts and random notes into blog posts soon.

Subscribe to the newsletter

Get updates about new posts, new projects, or other meaningful updates to this site delivered to your inbox. Alternatively, you can follow me on Twitter.