A Hard Look

I had millions of things I wanted to do. My main objective, like everyone else, is to make money so I can reach financial freedom. Thus, I had to ruthlessly cull certain things. Here were all the things I was doing/thinking about doing, what they are, and their pros/cons:

  • Quant Stuff: To algorithmically find alpha in the market, like Quant Quiver. Pros: Interesting. Cons: Extremely difficult and competitive and time-consuming. Likely to lose money. Thus, don’t do it.
  • Backrun Bot: To perform sandwich and backrun attacks on DeFI trades on Ethereum. This bot was up and running already, but is extremely competitive. For example, churning 1000 ETH in capital might yield 1 ETH in profits as the arena for DeFi arb is extremely competitive now. Too much time to maintain, for too little upside. Hence, abandon this months-long project.
  • Blog: fun, but very time-consuming. Little to no upside in revenue, especially for the stuff I talk about. Only do sparingly.
  • Doing research/shorting the market: Maybe do 1h/day research on weekdays to find good shorting or “put” opportunitites. Seems like low effort for relatively high money. As Jesse Livermore would say: “buy in a bull market, sell in a bear market.” It’s a bear market so it seems like easy money. Cons: Can lose a lot of money if reckless. Go for it, but don’t go too crazy.
  • SaaS: Build software and try to sell it. Financial upside: could be enormous. And unlike dropshipping/e-commerce, seems less competitive. Analyzing NAICS code would seem like SaaS is roughly 4X less competitive than, say, just marketing. Secondary upsides: I can use my projects as resume fodder and so even failed projects are useful for portfolio purposes. Tertiary upsides: Most software I want to build is to help my own side hustles. So even if I don’t get customers, I can automate some of my own workflow. Downside: Extremely time-intensive. Do this for 4h+/day on Mondays-Thursdays.
  • NFT arbitrage. I’ve been running this bot for a while. But lately there’s been more and more competitors, and very low trading volume. However, this has proven to be quite lucrative. So the upside is: I know this to be profitable, so keep maintaining the code/working on it. Downside: Getting more and more competitive over time, and any time OpenSea updates their API, everything breaks, so a lot of time is needed in maintaining the codebase. OpenSea also has notoriously bad documentation (i.e. missing information or flat-out wrong information). So that’s a pain. But it makes money, so do this 4h+/day on Fri-Sun.

Ultimately, it’s boiled down to just these 2 items: working on NFT arbitrage and a new SaaS company.

Rapid Learning: SaaS

I’m not a software engineer. Nor a web dev. So I had no clue what Django is nor how to make a website as of 3 weeks ago this time. In the past 3 weeks, I’ve:

  1. Took a week to run through a Django Tutorial, which was very helpful in helping me understand WTF Django is.
  2. Made an LLC.
  3. Started and seeded a business bank account.
  4. I’m in New York, so there are annoying publication requirements to meet as well. Luckily, there’s a service that does all this.
  5. Created and tested a website on localhost, where eventually I was able to push it to production on wraithscribe.com
    1. This is just a product that helps people rapidly write blogs by having them press a button.
    2. A lot of other services force the user to come up with all the blog sections before the AI would write for them. Wraith Scribe just does it in a single click.

Next steps: Find beta testers.

Business Architecture

Product-market fit is important. And marketing is arguably much more important than product. This is because marketing informs the product. Not all products should be on the market. Thus, my intensive schedule is roughly:

  1. 1.5 weeks to create a new SaaS product.
  2. 1.5 weeks to find and chat w/ beta testers. This helps me clean up the app/UI in preparation for step 3.
    1. If I cannot find any in 1.5 weeks, move onto next step anyway.
  3. 4 weeks to do intensive marketing and try to scale.
  4. At the end of 6 weeks, note down pros/cons/lessons and decide whether to scale or kill the product. Killing the product probably just means allowing only myself to use it.

Not So Rapid Learning: NFT Bot

As much progress as I’ve made on the web dev side in 3 weeks, I made almost no progress on NFT arbitrage. OpenSea moved their smart contract to a new address and whilst it was easy to generate correct encodings to call the contract “properly,” the transactions also only sometimes fail with no error messages. I haven’t made any forward progress on this as there’s no clues as to why things only sometimes fail. When there are failure messages, they look like this:

"revert": "��\u000ec�Fx�%����SH\u0001?BN#��\u0017^?\u0016r�<1n��/\u0006��"

"revert": "o~�&"

But it’s Monday now so I’ll visit this on Friday.

I wonder if NFT bot slowness is due to:

  1. Learning curve flattening so any progress is very slow progress.
  2. I’m incompetent.
  3. OpenSea just really, really sucks. If they actually bothered to put revert messages that were readable, development cycle might be much faster
  4. Or a combination of the above?

In Conclusion

Learning

If there is good materials out there, learning is quite doable. Take making a website for example. Plenty of information out there so it’s quite feasible to go from ‘zero to one’ in a matter of a few weeks of work.

Conversely, take the piece of shit enigma that is OpenSea. They’ll change their APIs and smart contracts without notifying you. You’ll just know as your bot stops sniping NFTs. Once you do realize you need to update your code, there’s little documentation. And the little documentation there is are factually incorrect.

Thus, if you want to pursue projects, it may be better to pursue projects with ample material. And forget about using tools with bad documentation or generated by an incompetent team (OpenSea, for example). You’ll save yourself the time and headache.

APIs and Libraries

Not all APIs are equal. And a lot of times, you will lose alpha with bad APIs. This is because of the enormous time opportunity costs it takes to ramp up. I wouldn’t be surprised if a SaaS project was making a ton of sales and its revenue surpasses my NFT bot just due to the speed of forward progress allowed in the former.

Conversely, bad APIs could serve as a good moat. Consider Uniswap API: very straightforward, but made ERC-20 arbitrage very ultra-competitive because everyone could understand it. So maybe it’s good that OpenSea is so hard to use.

So there seems to be a tug-o-war between the difficulty of using an API vs. its moat. But I think it’s bullshit. Things that are harder vacuously has their moat disappear as soon as they make their services better if that’s the theory. So your entire business strategy on building on top of something hard-to-use would be foolish for 2 reasons: 1) you’re building a moat controlled by an external factor, and 2) the moat shouldn’t come from a product’s difficulty because difficulty in and of itself provides no intrinsic value.

So my conclusion is: use tools/materials that are easy to use. Complexity != alpha. Let your moat come from your positioning and how you iterate with your customers.