Write the things, anon. Also what have you been doing lately

I clearly don’t have time to write endless long form content but people seem to want me to write things. And I suppose I have things to say.

The original vision of the Post Fiat Wallet – which is an AI agent that basically tells you what to do next and gives you rewards for doing so. Then over time qualifies your credibility and connects you with other individuals in a Hive Mind. As it turns out. Is quite complex to build. Even with AI agent coding tools.

So that is what I have been doing. Day in, and day out. Making a big update to our existing system that allows AI to augment your judgment and rewards you for doing things that are good for you. in addition to hiring ppl to do the things, and working w institutional clients (if you want to work for me and are a cracked C++ engineer, or a python AI master email [email protected])

Generally there’s a lot of hype when AI coding products come out. There’s always some guy who says “Look at these 96 websites that I shipped with ChatGPT using Codex” in the first 24 hours of launch. Then it’s inevitably som AI boot camp he’s selling. That’s the cynical side.

On the flip side, Claude Code is vastly improving the cadence at which you can ship things. Windsurf seems to have endless cascade errors. And Cursor requires too elaborate of instructions and hand holding re the parts of the codebase referenced. Whereas Claude Code is a happy medium between being able to do things that reference the entire code base, and also moving really fast.

Similarly, I’m increasingly a Claude 4 maxi both in terms of the chat capability as well as their alternative to the Responses API

The OpenAI responses API was originally quite nifty because it enabled both web search and vector embedding natively. You can pass standard vector objects directly into O3 and get somewhat coherent results. It makes demoware but when you get in the weeds, it is once again – pretty dicey

Good example - if you ask it for a particular Fed speaker’s opinion on inflation, the vectorDB has a hard time understanding the dates of different filings. So to make the entire thing work you need to spam dates through the data sets, or do fancy hacks to make it work. Which of course is going to limit its adoption.

It’s still decent though. And Claude web search and sycronous tool calls are pretty insane. Also – here’s one link you absolutely need to read (https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/claude-4-best-practices)[Claude 4 Best Practices]

I am sounding bearish but that’s just because I’m in the weeds actually doing things.

Great. What’s the plan, Stan?

But I guess the point of this blog is for us to answer the question concretely – how do you make it before AGI, and it’s a journal – and nobody has context of anything I’m working on

So I’ll outline my big bets / thematics and how I’m thinking about things

Advise the Ambient

My brother - Travis, runs a crypto AI start up called Ambient that I helped incubate and advise. It’s backed by A16z and Superscrypt. Travis ran AI R&D for a big agtech start up and his co-founder Max used to work with me at Perpetua and has been the founding engineer at a bunch of things (Blend, Cockroach Labs etc). Ambient simply put is AI bitcoin. That sounds audacious but if you follow TSMC and global fab capacity trends, basically everyone is heavily investing against AI data center capacity and despite the increasing hashpower of Bitcoin there isn’t nearly as much investment against BTC miner capacity. Which - if you think about it makes sense. BTC has a market cap which is 33% less than Nvidia. and Nvidia is just 1 company. AI is ultimately a lot bigger than crypto.

People in crypto tend to get this wrong and think that crypto is bigger than AI. So the Ambient bet is that AI driven proof of work can get more hashpower than BTC. And the mining subsidy that runs on the network allows Ambient, as a network to sell Deepseek R1 (which just launched a banger open source model) – at a discount. So unlike SOL which can have network fees from anywhere from <$1m a day to $20m a day, Ambient usage should be consistently extremely high. This creates a flyhweel where this means more hash power, and ecosystem products. And there are a whole lot of people who rightfully don’t trust model resellers so ‘verified inference’ as it calls, basically allows you to use AI / LLM in your work and prove you’re not being man-in-the-middled with your IP. Which is great for financial use cases.

So I advise on that. But that’s not most of my time. Most of my time is spent on Post Fiat.

Run the Post Fiat

At a very high level when I made Post Fiat I started with the question “If Fiat currency is going away, what do I want to keep my money in?”

And that is literally the biggest problem in society. If money is broken, then nothing is stable. Literally anything you build, like businesses - are essentially sand castles built on unstable foundations. And this thesis seemingly gets validated daily by the absurdity going on in Global Governments.

And when you really study currency, it’s not about shareholder returns. Currencies aren’t usually stocks. They are protocols for belief, ascribing value to different things in a fair (or at least objective manner). Fiat currency moves that authority to the government – which is obviously bad when it gets corrupt or prints a lot of money. And the Gold standard moves that authority to an arbitrary agreement that “Gold is the thing”. But Gold can get pretty nasty as well. If you study the history of money, the Spanish Empire vastly expanded its gold supply when it conquered South America. Got cocky. Then repeatedly defaulted on its obligations

The “Not your keys not your coins” meme. Was literally the basis of most historical sovereign defaults. So it’s a bit ironic – that now Michael Saylor, who has probably spent more time than anyone studying the history of money. Isn’t willing to verify his BTC holdings via proof of reserves. It’s almost like we are just doomed to repeat history in stupid, highly leveraged ways. Because schools are really bad and nobody learns anything. But I digress

So Post Fiat - is basically a new vision of currency. Currency as a coordination mechanism. And the rough idea is to take the most simple viable cryptocurrency that you can (XRP). Rip out the centralizing elements, and replace it with an AI based system to route value to people in accordance with how much value they produce. And because you can’t really bribe AIs, that system is pretty svelte. And most people already trust chatbots more than they trust their politicians.

Okay but what is Post Fiat and why should it exist

It all sounds pretty abstract but there are three levels to it.

  1. A blockchain AI system that defines all your personal tasks
  2. A blockchain AI system that defines all your financial decisions
  3. A blockchain AI system that determines the actual governance and rewards of the system

Over time - you build trust in 1&2 to give credibility to 3. So yes there are cash flows involved. But cash flows are not the main point. I don’t think Hyperliquid is every going to be the global reserve currency. Whereas if I build Post Fiat well, it’s more about creating a system that can accomodate higher order intelligence in a +ev tokenomic way

Typing it out, it honestly feels like I’m pumping some token on Twitter. Unfortunately, that is not the case. I actually believe this

And I think it starts with nihilism. I simply cannot look at the society we live in and say “This seems fine”. It’s so clearly not fine. None of the debt math makes any sense. And I know so many crypto whales running around Puerto Rico. I get to see the delusion of the ‘Sovereign Individuals’. Guys worth $300-400m who literally lock themselves up in their apartments all day for security reasons. Because society is dangerous.

BTC is an extremely antisocial movement. Its slogan is frequently, “Have fun staying poor”. And the entire pitch is “you’re already a peasant so you better buy or you’ll stay that way for all eternity”. I’m sorry. That is uninspiring trash that isn’t at all justified by the tech advancements on BTC. The lightning network hasn’t scaled. DiD hasn’t gained any traction despite much hype. People mostly use it as gold and a levered speculation instrument.

Don’t get me wrong. BTC is a lot better than simply having no plan and printing infinite money. But it’s not really a satisfying answer to “what do we replace currency with”

That is just a bunch of dudes, living in police states, with a lot of BTC, overpaying for things. While society basically rots.

So to me – Post Fiat really just started with an exploration of “how would the world not absolutely suck if fiat currency died”. And the answer inevitably was “AI systems need to integrate with currency directly and determine what is fair, because governments are failing at this”. Currency needs to embed commands, logic - etc

I actually think this logic is obvious and inescapable. But most people refuse to grasp it

But then– that’s super Zoomed out. Zoomed in, brass tacks, tactical. I’m making a cryptocurrency. It issues commands linked with money to people. We’re making a new L1 that is going to do this.

Cool story bro how do you make number go up

But I’m also a creature of markets – and I know people don’t really want a token where there’s going to be a ton of selling pressure.

Which brings us to the next question I’m navigating:

“How do you generate a ton of buy pressure on a cryptocurrency”

There are some higher order principles I’ve established to answer this question – namely

  1. Treat the currency as an e-commerce product and maximize its conversion rate across the full stack. That means doing non traditional things like launching merch, memeing on X, and running ad optimizaitons to know exactly who is buying my token, why and waht messaging resonates best
  2. Keep costs low. If people don’t work, fire them fast. Be frugal. Live in PR. Do smart tax things. Whatever it takes to sell the lowest number of tokens to see Post Fiat to conclusion as possible
  3. Consider alternative financing mechanisms such as ICOs where the holder base is more aligned with the project rather than accountable to fiat based LPs
  4. Create cash flowing enterprises bolted on to the currency from the get go. This brings me to my big “BD bet”

My big BD bet is that

  1. AI capital markets is where you want to be. Every single hedge fund and investment bank is all hands on deck re: integrating AI into its process
  2. I personally have extremely strong conviction that AI capital markets actually makes money so - A. launching treasury management products B. launching an ETF – the big difference with qualitative AI analysis is that it’s much more scalable C. building in compliance rails for crypto native expert networks where people get paid for expertise (think Tegus)

I am a big fan of ‘margin of safety’ thinking. And when you look at the macro: AI is very good at doing repeatable known process. But as far as we know hasn’t invented anything yet. And we have no idea if it ever will invent anything. Or if that’s 3 years away or 10 years away. But I can underwrite AI models consistently creating alpha trading – because I’ve seen it statistically significantly evolving over time.

I think the cool contrarian bet in AI capital markets is that the real edge in markets isn’t from analytics. It’s from coordination effects. In our community we have something called “the data lake”. I don’t know anyone’s name, but everyone is on Twitter. And has the capability of ramping narratives. Engaging different communities. So you start with a great analytics stack while we are in a rising tide and there’s still edge doing that. But over time the edge goes to zero. And by that time you have a strong, vetted community using the Post Fiat Task Node – who serve as the intelligence architecture for the alpha engine going forward. That is “Contributing to the Data Lake” or how to think about it – it’s the only real moat

Give me a summary you’re losing me

But yes that’s mostly the strategy.

But I wanted to just kind of put it out there even though it’s not perfect

  1. I am making an all in bet on fiat currency failing and being replaced by artificial intelligence
  2. I am doing this by running Post Fiat (and advising Ambient)
  3. Applying AI to capital markets is the first, best, and most conservative use case to put that bet on without selling too many tokens

I think where I’m contrarian is that I don’t think things like XRP are bubbles. I think they are indicators or smoke alarms. And that our job in the crypto space is to take absurdity and transmute it into life rafts for the coming fiat currency collapse.

Addendum: the people want an ETF

One final thought on branding – apparently 70% of my X audience wants me to launch an ETF. I have a pretty clean idea of how to do that so I want to spell out the ETF thesis here

Everyone is dystopian but there isn’t a good dystopia ETF. I think I can deliver that with the 3Ds

Debasement

  1. Long gold, silver, BTC, ETH, XRP, SOL or whatever spot / futures crypto product I can get on
  2. Short S&P 500 correlated FX crosses (AUDJPY, CADCHF etc)
  3. With an AI overlay – auditable – so you can see how AI is informing the weightings and decisions

Degeneracy

  1. Long exchanges, video games or other high margin addictive products with low customer acquisition costs
  2. Short the real economy

Delirium

  1. Long bubbles that are forming as measured by convergence of AI hallucination, narrative, and data tracking that are in uptrends
  2. Short collasping bubbles with opposite effects via put options

I’m going to look into what that takes and potentially use Post Fiat as indexing tech to publish the weights of the index in the meanwhile and provide them to task node members

End of Rant

Okay so apparently I have a lot to say. I’m going to keep ranting. Sometimes strategic, sometimes tactical. Try to keep a daily cadence. I think it’s important to document weird times in history, and how many times is fiat currency going to fail at the same time as human intelligence becomes obsolete.

Thanks for joining me on the Post Fiat Journey. This post was all over the place. I feel like one of those “Spider’s cobweb on caffeine” memes.

What I really want to do here though is help people understand what Post Fiat is, why I’m building it and the specifics of how I’m making this a thing. Even though it’s ugly, not poetic or hard to understand. Fuck it.