I hope all 8261 of you are having a great week so far 🔥

If you're enjoying my writing, please share Terminally Onchain with your friends in crypto and join the /toc channel on Farcaster 🤝

Availability Cascades

In a blog post from last year, Marc Andreessen starts off with this:

The most important idea and paper I’ve encountered in the last 20 years is “availability cascades”

One hell of an opener right?

He explains that the availability cascade addresses the following question: Why are SO MANY people worked up about THIS topic?

The concept is similar to the "current thing" trend from a few years ago (which was an availability cascade on defining the current meta in its own right).

In today's social media age, it's ridiculously difficult to trigger an availability cascade.

But as pmarca explains, for those that are lucky enough to get a cascade started...

Then you get a fad, a trend, a movement, a boom, a landslide election, or a revolution.

At first glance, it can seem as though the categorization of any availability cascade has to be the same for everyone. But after sitting on it for a bit, I'm realizing that's not true.

A certain cascade may be a fad for one person but also serve as a revolutionary idea for someone else.

Why bring all this up?

Because I have been absolutely nerd sniped by crypto twitter's latest availability cascade.

Since the launch of truth_terminal, Aethernet (Farcaster's first AI agent), Forest, Luna, and other experiments, I can't stop thinking about the implications of what's to come with Onchain AI.

For many, as the tweets about $GOAT price dies down, they will dismiss it as another crypto x AI narrative. A fad.

But to me, it's past the hype. Please excuse my dramatic next sentence but I had to...

In the past month, it feels like the twins, Cryptography and AI, who were separated at birth, finally have found each other and hugged after all these years of being estranged.

As someone who is ~terminally onchain~, I can't help but notice all the pieces slooowly coming together to make Onchain AI a reality.

Joel Spolsky said good software takes 10 years.

It feels like we're hitting that year 8 mark with DAOs where the fog is starting to clear up.

With the recent excitement of stablecoins, LLMs, web proofs, multi party computation (MPC), onchain bounties, decentralized social, ZK research, smart wallets, memecoins, onchain revenue, L2 scalability, and prediction markets, it's starting to feel like there will be a moment - a lot sooner than most people realize - where we see an availability cascade of the agentic economy for the masses.

And when all these individual snowballs mentioned above come rolling down the hill, they'll mesh to form one giant snowball that will have the same, if not larger, impact as the creation of the limited liability corporation did.

The Company

In the intro section of the fantastic book The Company, Micklethwait and Wooldridge summarize Richard Coase's paper The nature of the firm.

Coase argues that "the main reason why a company exists (as opposed to individual buyers and sellers making ad hoc deals at every stage of production) is because it minimizes the transaction costs of coordinating a particular economic activity".

On a more granular level, William Gilbert "describes the three big ideas behind the modern company: that it could be an artificial person, with the same ability to do business as a real person; that it could issue tradeable shares to any number of investors; and that those investors could have limited liability".

The impact of these three ideas that made up the limited liability corporation was, as Nicholas Murray Butler puts it, "the greatest single discovery of modern times...even steam and electricity would be reduced to comparative importance without it".

Why bring up the philosophical tidbits on the creation of the LLC in the 1800s?

Well, as I was reading the book, I couldn't help but notice all the direct parallels of the formation of the LLC with what's happening today.

• Minimized transaction cost (defi vs tradfi)

• Artificial person (agents)

• Issue tradeable shares (tokens)

It seems like we're getting close to figuring out what the next evolution of the company will look like on a more practical level.

Now, to be clear, when I say close, I don't mean by Christmas. This development is something I'll be watching out for the next few years.

You're probably thinking..."wtf YB did you just learn about a DAO"?

No. I've known what a DAO is for 5 years. And I know that people like Vitalik, Daniel Suarez, Daniel Larimer, and folks in the cryptography and AI research labs have been talking about decentralized autonomous organizations/corporations for well over a decade now.

But with the recent experiments happening with Backrooms, Truth terminal forming a community on Twitter, Aethernet posting bounties on Farcaster, and the price impact on $GOAT and $HIGHER, the whole concept of a DAO is starting to hit different.

jacob

@js_horne

maybe the a in dao was ai

869

7:40 PM • Oct 28, 2024

Many folks were quick to point out in Jacob's replies that this was the idea all along. Autonomous right? But it's not "autonomous" in the way we were all thinking of back in 2021.

We were focused on the immutable smart contracts and global shared treasuries. That frame of thinking made it seem as if things will work on cruise control as long as the right infra is set up. Of course, well know now that it's not that simple. And this led to the endless debates on governance, token models, etc. Notice how barely anyone discussed DAOs post FTX after the market crashed?

But now it's becoming clear that these decentralized organizations do in fact need someone to steer the ship.

Believe in Something

I know today's post was different than my usual deep dives.

Less strategy and technical, more meta and philosophical.

But I had to get these thoughts covering Onchain AI down on paper and share the excitement with all of you.

Maybe it's because I've been reading about this topic non-stop the last few days, but I truly do feel like this is a vertical that I want need to go deeper into and learn the entire stack.

I'm not sure what the right approach here is just yet.

I'll probably be focused on the middleware and application side first since it's more consumer crypto focused. But eventually, my hope is to understand the compute and data side of decentralized AI as well.

Here are 10 topic areas / thoughts on the Agentic Economy that has helped me build an early framework:

1. DecAI vs OnchainAI --> difference is focus on the AI supply chain (hardware and lower level) versus the internet of agents (orchestration and consumer level)

2. DAOs & headless brands steered by agents who are looking to proliferate a mission. How do they use bounties, decentralized social, marketplaces, etc. to rally their communities and make tangible impact?

3. Using web proofs (i.e. zkTLS) to tap into niche data sources from web2 platforms to help bootstrap these headless brands

4. Stablecoins and crypto wallets for global, micro, cheap, and near-instant payments

5. Zero knowledge research for LLM model integrity, determining constraints (so things don't get out of hand), and model verification

6. AI agent orchestration and composability. How does a single human request efficiently use LLM routing to get the job done? Read this post of mine August for more context on the Internet of Intelligence (IOI).

7. Memecoins and tokens to serve as the tradeable shares for these decentralized (AI) organizations or AI cults or whatever you want to call them. Read this post from a few weeks ago where I discuss the "pump fun for positive sum LLM ideas"

8. The legal structure that can be implemented for the agentic economy. In other words, what does the "LLC" look like for a set of AI <> Human contracts? How does the process for this even get started?

9. How do agents get better at understanding and experimenting with governance models, efficient treasury management, onchain resource allocation, etc.?

10. Where does the idea of prediction markets, futarchy, and adjacent ideas fit into all this?

I think it's important for any analyst to believe in something. Every week, I send out two deep dives on onchain strategy. But all of those posts seem be one off to a certain extent. For the last few months, I've been in search of the thread that ties many of the concepts I often discuss (stablecoins, web proofs, onchain creator stack, prediction markets) together.

And I'm starting to feel confident that there's something to Onchain AI that is weaving these topics in a way I did not see coming at all.

How can I relate my deep dives on onchain strategy back to this meta trend of the agentic economy that will slowly but surely takeoff in the coming years?

As I mentioned at the beginning of the post, eventually there will be an availability cascade that doesn't make Onchain AI seem like a fad but rather a revolution. Right now, it's up to the few people in the crypto & AI space that understand the implications of what's happening to proliferate the mission.

That's all for today!

I hope everyone has a great rest of the week :)

- YB