Update 1: GPT-3 Idea Generators & The Interface Layer of Web3

Thinking about how to design a more useful and usable web3.

May 25, 2022

Anddddddd it’s newsletter time! Thanks for taking the time to sign up and read.

One of my goals for this year is to be more public with things I’m interested in and thinking about. My hope that in writing publicly I can better structure my own thoughts, and hopefully share something interesting with you as well.

A few months ago I left my role at Clay, the startup I co-founded, in order to explore web3 full-time. I found myself increasingly drawn to the web3 space and curious about the possibilities for the types of future software services and organizations that could be built - enough so that I decided it was time to take the leap.

I took a a few months off for rest, which I spent mostly learning to surf. After more time off than I thought I needed, I’ve started started to get back to work: learning, investing, and building in the areas that I find interesting in the web3 space.

I’ll hope to send these updates out at least bi-weekly. The focus of my time and writing will largely be web3, but I’ll include other things that I’m also finding interesting in tech or beyond.

Personal Update

I’m in Aarhus, the second largest town in Denmark, visiting my sister who’s a medical researcher out here. It’s been a lovely last few weeks getting to spend more time with my niece and nephew. I rented an AirBnB with a view of the ocean, and have been making the best of the long summer days: the sun rises before 5am and sets close to 10pm each day. It’s been a good space to start diving into work!

I’ve been getting in lots of long runs in the forests here, and been loving the Danish coffee and food - lots of bread, butter & cheese from the local cafes :)

Things I worked on these past two weeks

Possible. I shipped a fun little side-project experiment I’m calling Possible (possible.vision). It’s a GPT-3 Idea Generator that sits on top of AngelList data.

Possible is a way to explore possible ideas by smashing two market categories together and asking OpenAI’s GPT-3 to describe possible ideas in that intersection. It was a good chance to play with the OpenAI GPT-3 services which I’m consistently impressed by.

You can read an in-depth post about why I built Possible & related thoughts.

Writing. I’m working on a piece around re-designing how we handle and debug errors in software design.

Web3: Learning, Exploring & Building

I’ve been debating how much existing knowledge of web3 & crypto to assume when writing these updates. To keep it interesting for folks following the space more closely, I’m going to assume a somewhat advanced knowledge of web3 and the related terminology.

I’ll do my best to help set up context and give sufficient background to still make it broadly approachable for people wanting to follow the threads. If there’s anything I write about you’d like me to unpack more in future posts please just send me that feedback.

Where are we in web3? What’s getting me excited?

I believe that the most interesting applications of web3 are all yet to come, and right now we’re in a phase where the limitations of the platforms, both on the scaling side, and the UX side, are preventing the most interesting services from being built.

But, with L2 scaling solutions on the way, new types of services will start shipping, and despite the current bear market we’ve entered, I’m mindfully bullish on what’s next. I do think there are a lot of very valid concerns to pay attention to as the space moves forward, but I also think there is also a lot of potential. Even some of the launches from these past few weeks have me excited for whats to come.

This first update I’m going to use as a background on some high-level themes I’m thinking about, and using future updates for more focused posts. I’ve kicked off my deep dives by reflecting on how to bring better design to web3 in order to make it more usable, understandable, and useful.

The main themes I’m interested in exploring these next few weeks are:

How to improve the interface layer of web3

When we think about how we interact with web3, there’s a few different pieces that make up our “interface” into web3. Across both desktop & mobile we have:

our browsers
wallets (as apps or chrome extensions)
application UIs powered by web3 smart contracts and infrastructure under the hood
block explorers like Etherscan for browsing smart contract code & on-chain state
Github repos that host the code for the corresponding smart contracts

Collectively I’d say these make up our “interface” to web3, and I’m thinking about how each parts of these might be designed to make a more intuitive and compelling experience of web3 services. These questions apply to both the end-user and developer web3 experience, and they will be additionally relevant as the space evolves over the coming years to navigate the growing complexities of a multi-chain world.

More concretely, there are two areas I’m exploring under this theme of better web3 interfaces in multi-chain worlds.

A Web3 browser/wallet interface and security experience that everyone can feel comfortable using. How do we make the experience of using wallets feel much more intuitive and secure than it does today? There’s a potential upcoming shift to more easily allow for wallets as Smart Contracts on Ethereum. This could change how wallets work, and allow for a more “controller”, “user-friendly” experience
Etherscan is hard to understand. How do we build more intuitive interfaces for exploring and understanding on-chain data and smart-contract systems? If you’ve tried to follow a smart-contract system like ENS or Livepeer using Etherscan or their other chain explorers like Polygonscan, it’s very challenging. As we enter a multi-chain world, where services tend to span multiple smart-contracts, we need a better tool-kit for navigating the complex systems and on-chain data we’re setting up.

Web3 Everywhere & more useful applications.

How does web3 start to get much broader adoption and go main-stream in obviously useful applications of the tech? I think one of the criticisms of web3 is that it’s not doing anything “useful”. Just the other day Bill Gates in a Reddit AMA said he doesn’t own any crypto because:

I don't own any. I like investing in things that have valuable output. The value of companies is based on how they make great products. The value of crypto is just what some other person decides someone else will pay for it so not adding to society like other investments.

Supporters of web3 trying to counter Bill Gates could point to DeFi (setting up a programmable, open, international financial system), or NFTs as compelling initial use cases, but for most people it might still not be appearing in their day-to-day lives in any way that feels like it’s making a significant useful impact. I believe we’re starting to see the early-innings of unequivocally more useful web3 services, and I’m spending time thinking about what it looks like for web3 to be broadly “embedded” and used by everyone in obviously useful ways (the operative word being obviously, i.e. not needing a long explanation about long-term visions of shifting societal structures). I think web3 social, decentralized compute services, and public goods are some examples that I’ll dive into over the coming weeks.

Embedded Web3: Public goods, and a modular plugin system for the web.

Web3 changes how we build software services, and the types of software that can be built.

The thing that gets me perhaps most excited about web3 is that it opens up the possibility for funding and building services in ways that weren’t possible before (particularly for software services).

An older post from 2019 I really enjoyed around this topic is “What Comes After Open Source” by Denis Nazarov (@literature), who’s went on to cofound Mirror.

The combination of novel incentive and funding mechanisms changes who builds software, and how software gets built. I think organizations (DAOs) that figure out how to tap into these incentives will have new approaches to funding and building these new types of software.

Furthermore, the combination of everything living on a unified “world computer”, where people go around from app to app with their identity, data, and payment services in hand (via their wallet) leads to portable, modular, inter-operable software. This changes what parts of software the developers need to spend their time re-building. As things like identity, social graphs, payments, and data integrations are already taken care of by decentralized service providers, developers can then focus more time on other layers integral to their applications.

Beyond web3, I spend time consuming content from the future of computing / tools for thought community, which gives me another lens on the space. The intersection of web3 and the future of programming is one area that I’m excited to explore.

I recently listened to this episode of Metamuse, a podcast hosted by the future of computing research lab Ink & Switch. I’m a big fan of their work. In this episode they were chatting with Molly Mielke (mollymielke.com) about her writing on building better, interoperable, creative computational tools. I really like Molly’s work too, and would recommend her piece “Computers and Creativity”.

Some of the big questions that the creative computation community are exploring is how to build interoperable and moldable software. For example, Molly ends her piece on creative computation by writing

Interoperable, moldable, efficient, and community-driven digital creative tools hold immeasurable potential as co-creators with human beings. Tools of this type would lower the barrier to entry and make all users toolmakers and owners in an expanded definition of technological innovation.

I think web3 has the potential to be a big part of this interoperable, moldable future. Lens Protocol which I discuss below is one example of the new types of interoperable services that can be built. Trying to imagine how these new, portable services will surface in the interfaces we use is perhaps the area that most excites me.

Launches & Interesting Things I Learned

Coinbase as the Interface Layer for web3

Security & wallet UX challenges are part of the reason that many people would rather have Coinbase manage their funds and wallet keys. For the most part, because there’s not that many “useful” applications anyways, people haven’t yet been beating the doors down for wallet experiences that make it super easy to have frequent interactions with dApps.

One interesting developments from these past weeks that I think is worth tracking on this front is that Coinbase Wallet launches a built-in dApp explorer

Their intro post was titled ”Access web3 with the Coinbase app“. I think their approach feels very similar to the AOL approach of the 1990s where AOL tried to be the walled-garden that contained the entire “web”, by bundling dial-up, plus their own web portal, e-mail and messaging service. “In 1997, about half of all U.S. homes with Internet access had it through AOL”. If Coinbase is AOL, trying to be the walled-garden of web3, what’s the Netscape approach?

AOL was both a “dial-up” / broadband internet connection & then tried to also be the software layer via their siloed web portal, eventually acquiring NetScape for $4.2bn to be able to play in the more open browser space (important to remember that “a” of “a16z” was the founder of Netscape and knows these lessons well :) ). They lost to Mircrosoft’s larger installed user-base and also to the rise of broadband internet - perhaps competing on too many offerings at the same time.

Lens Protocol: web3 Social

My favorite web3 development over the past few weeks was the launch of Lens Protocol (lens.dev) on to the Polygon Mainnet. You can read about it more in their launch blog post “Web3 Social is Blooming: Lens Launches to Polygon Mainnet”

Lens Protocol is a web3 social networking protocol. Imagine an extendible Twitter or Facebook, that was not owned by a centralized corporation, but instead was a protocol that could be easily used as the foundational layer for building new versions of Reddit, Twitter or Facebook.
The Lens Protocol gives some very simple primitives like creating profiles, following other profiles (every follow is an NFT), and posting/collecting content (also NFTs). The data stored on-chain can then be brought to any application - so you could for example bring your contact book to any app you go to, so that each application doesn’t need to re-create your entire social graph.
You can think about it almost like a decentralized version of what Facebook Connect tried to do back in the day (remember the sign in with Facebook and see your friends in different apps functionality that was big in the early 2010s?) - hopefully it doesn’t turn out like that though, but we should keep in mind why those past attempts from centralized social networks failed. I think a big part of that might be figuring out how to leverage the protocol to have different social graphs for different applications/use-cases.
The Lens project is what got me re-excited about web3 a few months back. In particular I got excited about their really interesting extensibility model that makes it so that what a “Follow”, or “Collect” (save/like) does can be expanded based on the applications needs - you can read about their Modules system for extensibility.
It’s interesting to imagine how you’ll be able to embed Lens social graphs and social experiences into any application. Going back to the earlier points about public goods, and interoperable software - Lens Protocol and the ecosystem of apps already being built is a great example of what’s to come.

Miscellenia

I also listened to a few Bankless DAO Podcast episodes I liked, here’s my one line take-away for each:

How to be Early | Olaf Carlson of Polychain. It’s going to be really hard to predict how everything will play out in the L1/L2 space and where value will accrue, so it’s best to stay active and participate in many eco-systems. It seems like an index ETF of L2s and L1s would be an interesting bundle to set up.

The Trillion Dollar L2 Opportunity. I really liked the language they used to refer to all the L2s building on Ethereum: “The United Chains of Ethereum” - and how they’re essentially like States in the US - groups that get to piggy back on the security system of the nation: Ethereum. My takeaway was that the dynamics of L2s are such that they could get adoption very quickly and disintermediate users from directly interfacing with Ethereum as trust in them compounds. It’s good to start trying out the UX of these various protocols, and their corresponding wallet dynamics.

A line that I’ve been thinking about that popped into my head the other week was:

“How do we prevent the internet of money from becoming the internet for people who have money?”

On the non-tech side I’m reading “The Selected Works by Audre Lorde” and really enjoying it. One passage from the first essay in the book “Poetry is not a Luxury” which I really like:

As they become known to and accepted by us, our feelings and the honest exploration of them become sanctuaries and spawning grounds for the most radical and daring of ideas. They become a safe-house for that difference so necessary to change and the conceptualization of any meaningful action… We can train ourselves to respect our feelings and to transpose them into a language so they can be shared. And where that language does not yet exist, it is our poetry which helps to fashion it. Poetry is not only dream and vision; it is the skeleton architecture of our lives. It lays the foundations for a future of change, a bridge across our fears of what has never been before.

Thanks & Feedback

Hope you found something interesting in all of this, and let me know what else you’d like to read more about! Future updates & posts are likely to be more concentrated around one particular topic for the week.

Nicolae’s Newsletter