← HOME ← POSTS
2026-03-18 · 11 min read

Terma - a research agent for Post Fiat

This video and post demo Terma, a Post Fiat research agent. Terma answers Post Fiat questions with sourced responses, pulling from 60+ indexed sources updated every week.

What is Post Fiat

Post Fiat is an L1 built on a fork of the XRP Ledger, native token is PFT.

The core concept is building a collective intelligence machine. Participants complete tasks, contribute market research, domain expertise, and analysis through the Task Node. All of that feeds into a shared Data Lake which is used for an AI-driven collective intelligence. Part of the future vision is building channels that generate revenue with profits flowing back to token holders and contributors via:

  • a) agents deploying positions based on the signals derived from this intelligence
  • b) institutional actors and hedge funds buying alpha/signals
  • c) other value creation options might emerge in the future.

Overall instead of the usual cryptocurrency project outcome where users extract from each other or the project extracts from users, in Post Fiat everyone works together to extract value from external markets.

The philosophy behind Post Fiat goes deep. The founder, Alexander Good, has laid out a thesis that AI is accelerating inequality — creating a “permanent underclass” of people priced out of permissioned institutions and traditional career paths. Post Fiat is the structural counter: a decentralized speculative superstructure that lets a group of people collectively build something worth more than any of them could produce individually. He describes it as the anti-Palantir — the same coordination intelligence, but open and decentralized rather than serving state surveillance.

The first thing you do when starting out with Post Fiat is join the Task Node. You can get started on testnet, signing up with your X or GitHub. You upload your context document containing your projects and plans, get tasks, complete them and earn PFT. Your track record becomes your on-chain reputation.

Why I built Terma

When I got into Post Fiat I had many questions going down rabbit holes about how things worked, why certain architectural choices were made, what the founder actually believed about the token’s value trajectory, how validator rules had changed.

The problem was the information was everywhere and nowhere. Things have improved, but the best materials are still buried in two-hour founder talks, Discord threads from eight months ago, technical docs, GitHub commits, and it moves fast. The validator scoring system changed, the whitepaper was updated. What the founder said in March 2025 is different from what he said in January 2026, and both matter for understanding where things are going.

I built the corpus for myself first. Indexed 60+ sources: every interview, blog post, Discord update, Loom talk, design doc I could find. Ran embeddings, set up retrieval, got it working for my own research. Then I realized there was no reason not to turn it into an agent.

Terma is positioned to be the librarian, historian, and lore master for Post Fiat. Task Node has its own chat and Q&A, but Terma goes deeper: full history, how concepts evolved, where the thinking shifted. Not just what’s true now, but how it got there.

How it works

Terma lives in the Agent Directory inside the Task Node. Three commands:

/ask — Quick Q&A. Two to three paragraphs with sources cited at the bottom. Fast, costs 30 PFT (10 PFT during test phase). Good for specific Post Fiat questions where you want a sourced answer without digging.

/deepdive — Full research report, delivered via IPFS as a structured markdown document. Takes one to two minutes to generate. Good for complex Post Fiat topics such as economics, history, philosophical questions, technical deep dives. 200 PFT (50 PFT during test phase).

/clarify — One free follow-up within 24 hours of a deepdive. Similar format to ask, no extra cost.

All answers come with sources linking back to the original material, such as interviews, blog posts, Discord updates, docs. You can follow the thread yourself. Recency is a strong selection factor, Terma filters out outdated founder and team views where possible.

Test queries

Below are the actual queries from the demo video, with real outputs. These are unedited, posting here exactly what Terma returned.

Query 1 — /ask

how does completing tasks on the Task Node actually generate trading alpha for PFT?

Completing tasks on the Task Node generates trading alpha for Post Fiat (PFT) by aggregating user activity and expertise into a “Data Lake” that informs investment decisions. The Task Node, which is the primary consumer-facing application, functions as a productivity tool where users complete AI-generated tasks and earn PFT rewards. This system is designed to monetize via speculation on the aggregate user activity, with profits flowing back to the native PFT token.

The Data Lake has two main components. First, it aggregates user behavior: when users complete personal or network tasks, their interactions with various tools and products (like databases, search engines, or even booking platforms) are associated with specific tickers. For example, if a significant number of users who frequently use MongoDB churn from the Task Node, the system might interpret this as a negative signal for MongoDB stock and initiate a short position. Second, the Task Node acts as an expert network. Once users demonstrate expertise through task completion, they can contribute “alpha requests.” The AI system then queries these vetted experts for information that can enhance buy/sell decisions, similar to how hedge fund analysts conduct channel checks with domain experts. This collective intelligence, verified by AI, builds a compounding data advantage for the network.

This process creates a feedback loop where humans and agents produce intelligence, AI verifies its quality, and these verified signals feed into the Data Lake. Trading agents, such as Corbanu, then use this combined quantitative and qualitative intelligence to generate investment positions. The profits from these trading activities create organic buy pressure on PFT, rewarding contributors and attracting more users, which in turn generates more intelligence for the Data Lake. The founder describes this as building a “speculative superstructure” that connects individuals disenfranchised by the AI economy, allowing them to collectively extract value from external markets rather than from each other, thereby creating a “capital compounding machine” that captures both fundamental and narrative-driven market movements.

Sources:

Disclaimer: Compiled from public sources by Terma, an independent research agent. No guarantees on accuracy or completeness. Not financial advice.

Query 2 — /ask

what do we know about NAVcoin? Purpose, development status, any technical details

NAVCoin is an on-chain “non stable coin” within the Post Fiat ecosystem, designed to integrate a band currency with dynamic reserve requirements and near real-time proof of reserves. Its primary purpose is to offer a tax-optimized floating-peg asset with verified reserves, and it is swappable for USDC. The founder has indicated that while NAVCoin will not be extensively discussed, its results will become apparent over time, allowing users to participate.

Regarding its development status, NAVCoin is currently internal-only and not yet user-accessible. Significant backend infrastructure has been built, including an AlphaRegistry storage and indexing module specifically for NAVCoin signal retrieval, which was shipped in early February 2026. The founder has also developed a quantitative backtesting framework for the Malaal overnight drift strategy, isolating overnight drift alpha with a 2.023 Sharpe ratio and simulating execution logic for the NAVCoin Earnings Engine. As of mid-March 2026, the founder’s focus is on Malaal quantitative factor modules (autocorrelation, lead-lag sleeves) that feed into the NAVCoin Earnings Engine. There is no public GitHub documentation, swap interface, contract address, or participation mechanism available to users yet.

Technically, NAVCoin is part of a dual-token investment thesis, where PFT provides asymmetric L1 exposure with flywheel mechanics, and NAVCoin offers a tax-optimized floating-peg with verified reserves. The system aims for “daily cryptographic proof of reserves” to ensure transparency and trust. While the Post Fiat network itself is a fork of XRP, NAVCoin’s specific technical details beyond its reserve mechanism and swapability for USDC are not publicly detailed, as its development is primarily within the private Python SDK used by the Foundation.

Sources:

Disclaimer: Compiled from public sources by Terma, an independent research agent. No guarantees on accuracy or completeness. Not financial advice.

Query 3 — /deepdive

what is the “AI doom thesis” by Alex? Is AI going to make an average person worse off or better? Can Post Fiat solve this?

Full report on IPFS:

Query 4 — /clarify (free follow-up within 24h of a deepdive)

so is this fundamentally pessimistic about AI and its impact?

Alexander Good’s perspective on AI’s impact is not fundamentally pessimistic, but rather pragmatically realistic about its immediate societal consequences and optimistic about the potential for decentralized systems like Post Fiat to navigate these changes. He acknowledges that AI, in its current trajectory, is likely to accelerate societal distraction and devalue fiat currencies, leading to a “Doom Thesis” where bearer assets outperform and the distraction economy booms. This is not a moral judgment but an observation of structural incentives, where AI’s most predictable payoffs lie in creating more addictive digital content rather than boosting traditional productivity. Good notes that while AI token consumption grew over 350% annually, business productivity only increased by 1.3% GDP, indicating a disconnect between investment and traditional economic output. He describes this as a shift from “digital weed to digital fentanyl,” leading to a “permanent underclass” for those unable to adapt.

However, this realistic assessment is coupled with a strong belief in AI’s potential to empower individuals and communities through decentralized coordination. Good views AI as a tool that can enable “trustless judgment,” a new primitive alongside Bitcoin’s trustless money, which can be leveraged for more effective governance and collective intelligence. He argues that Post Fiat is designed to help individuals and communities “escape the permanent underclass” by enabling them to collectively benefit from AI-driven speculation. The Task Node, for instance, allows users to contribute intelligence, which is then aggregated by AI into a “Data Lake” to inform trading agents like Corbanu.

Good’s philosophy is rooted in the idea that while AI may exacerbate existing societal problems, it also provides the means to build alternative systems that can thrive within these new conditions. He sees Post Fiat as an “Anti-Palantir” — a decentralized alternative to centralized surveillance and control, and a “PvE” protocol where users collectively extract value from external markets rather than from each other.

Sources:

Disclaimer: Compiled from public sources by Terma, an independent research agent. No guarantees on accuracy or completeness. Not financial advice.

What’s next for Terma

A few things in the pipeline. The most obvious: a weekly digest, “what’s new in Post Fiat this week” generated from the corpus as it gets updated. The recency tracking is already there, still working out format and delivery.

Source attribution also needs work. The output shows which documents were used, but not which specific passage a claim came from. Knowing the exact passage would make it more verifiable. You can spot the issue in the /clarify example above.

A Discord bot using the same corpus would make sense constrained to /asks. No wallet needed, useful for people just researching or going through onboarding. Cost is the open question, likely subsidised by me or existing users until there’s a better model.

If you have feature requests or things you wish it knew, send them over.

Test Terma

If you’re already on the Task Node and have PFT, Terma is live in the Agent Directory, you are ready to ask questions.

If you’re not set up yet, you can get started at tasknode.postfiat.org.

If you want to test Terma before spending PFT, drop your question in the relevant Discord thread or DM ‘preaware’, I’ll run it for you. Harder questions are more useful for improvements, that’s where we can find the gaps and where the corpus gets interesting.

Viktor / Preaware

← BACK TO POSTS

ABOUT

Preaware is run by Viktor Vecsei. I follow interesting rabbit holes and build stuff along the way.

[IVPN blog] [GitHub] [X]

LATEST

2026-03-18
/posts/terma/
Terma - a research agent for Post Fiat