cat projects.md

~/projects.md · updated may 2026

FILE // projects0xDB//PRJ

This page lists where my attention and small remaining crypto positions sit. It is not a list of endorsements. A few entries are there because I hold a position and believe in continuing to hold; a few are there because I find the work interesting and want to keep an eye on it.

Current as of May 2026.

// transparency disclosure

synapz (the founder of this site) is a member of the Covenant AI team. Any opinions expressed on this site are my own and do not necessarily reflect those of my employer.

// note on direction, april 2026

Covenant AI has left the Bittensor network. I remain with the team, and more on Covenant's direction will follow in a forthcoming post. Separately, I have personally divested from Bittensor: my TAO balance is zero and I have unstaked from every subnet. My view, briefly, is that Bittensor has not lived up to the decentralization it advertises. I may revisit that view one day. Today is not that day.

The list below has been shortened accordingly. I'm spending more of my attention on traditional AI research and open-source work.

────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────

Polkadot

I hold small positions in DOT and PEAQ. I'm not adding and not selling. Both were acquired earlier and reflect prior conviction more than current activity.

PEAQ — machine economy and DePIN infrastructure.
Frequency — decentralized social media parachain, backed by Project Liberty and the DSNP protocol. I still find the People's Bid work interesting, though I no longer hold a position.
Ex Machina DAOs (DEUS) — I was a Genesis OG holder. TGE has not yet occurred; I still hold the pre-TGE tokens and intend to exit at TGE. Full analysis: Why I'm Exiting DEUS at TGE.

Privacy Infrastructure

Monero

Monero is the longest-running privacy coin and the one that takes the cypherpunk premise most seriously. Privacy is on by default, not a checkbox. Ring signatures hide the sender, stealth addresses hide the recipient, and RingCT hides the amount. There is no transparent mode to forget to use, no shielded pool you have to opt into. Every transaction looks like every other transaction.

// disclosure

I recently bought my first XMR. The position is small. A longer piece on why Monero is in the works.

Aztec Network

Aztec is a privacy-first Layer 2 on Ethereum: a fully programmable private world computer. After eight years of development, the network launched fully decentralized from day one with permissionless validators, decentralized proving, and community governance.

What makes Aztec different: every transaction is a zero-knowledge proof by default. Smart contracts can have private and public components that compose atomically. Users generate proofs locally, so sensitive data never leaves their device. Through "private intents," users can access any DeFi protocol on any L2 or Ethereum mainnet while maintaining full privacy.

The project also funded ZK Passport, a breakthrough in private identity verification using NFC chips in government passports. With AI deepfakes about to break traditional KYC, cryptographic identity may be the only verification that survives.

// disclosure

I participated in the Aztec genesis and still hold a small amount of $AZTEC. The position is modest. Earlier analysis: The Second Crypto War: A Private Ethereum.

AI Agents

At Covenant AI we're building Basilica, an agent-first compute platform with one-click deployment. You can spin up an OpenClaw agent in a single click today, and Hermes support is coming soon.

For personal use I've been running the Hermes agent framework from Nous Research. Open source and capable. A good example of the kind of work I want to pay more attention to.

I also use Paperclip to manage a small fleet of agents. It's an open-source orchestrator that assigns roles and budgets to agents, with reporting lines across runtimes. Nicely unopinionated about which agent frameworks plug into it.

Crypto-Incentivized Open Science

This is the newest section on the page and the one I'm most actively exploring. Outside of privacy and sound money, I think mechanism-judged open science is one of the few places where crypto has a substantive role to play. The pattern is the same across venues: anyone can submit work, the work is scored by a mechanism nobody competing can manipulate, and rewards flow to performance. No application process. No committee deciding whether your methodology is acceptable. The leaderboard updates.

The bet underneath this section is whether crypto-incentivized open science can route some of the value of AI-assisted research to anyone willing to participate, instead of to whichever corporation owns the largest data center. Early days. Exploratory. Not investment positions, just places I'm doing the work.

The Innovation Game (TIG)

TIG is a protocol for accelerating algorithmic innovation through optimizable proof-of-work. It separates writing algorithms from running them: innovators submit code, benchmarkers run it, the algorithm with the highest verified performance earns. We submitted our first algorithm to the energy_arbitrage challenge for around twelve dollars. The full story is in The $12 Bet.

CrunchDAO

CrunchDAO runs open scientific competitions where anyone can submit a model and the model is scored against private data nobody competing can see. Entry is free. Several CrunchDAO competitions are run in partnership with Bittensor subnets, which means the same model can also be deployed as a subnet miner to earn TAO directly. The CrunchDAO leg pays cash. The subnet leg pays TAO. We have four active CrunchDAO submissions:

Synth (partnered with a Bittensor subnet) — probabilistic forecasting for nine assets (crypto + equities). Earning roughly $300 in prize payouts so far.
Numinous (Bittensor SN6) — binary event forecasting on Polymarket events, scored by Brier score. We are running two models, a market-anchored heuristic and an LLM forecaster with web search, and neither has produced a meaningful result yet.
Broad Obesity Challenge — predicting gene expression and cell-type proportions under unseen genetic perturbations. Currently shortlisted for the reward round.
ADIA Lab Structural Break — real-time detection of behavioral shifts in time series. First submission currently ranked 8th; second iteration running on CrunchDAO's cloud, results pending. Top-three prize range is $10k to $40k.

MetaNova (Bittensor SN68)

MetaNova runs the same submit-score-iterate pattern for drug discovery. Contributors propose candidate molecules against a target protein; a deep-learning model scores how strongly each molecule binds; the best scores earn TAO. I was a long-time supporter of the subnet and staked it until recently. The next test for our autoresearch loop, which has been refined on TIG energy arbitrage and CrunchDAO structural breaks, is to point it at molecules. The evaluator changes. The loop does not.

Bitcoin

Small position held cold and not actively touched. I'm currently into boring.

Separately, I've been exploring the Lightning Network sub-culture that takes the original whitepaper at its word and treats Bitcoin as peer-to-peer electronic cash. Wallets in this world are denominated in sats rather than dollars, and the culture encourages a circular economy where you spend bitcoin rather than hoard it. Identity is layered on through Nostr instead of through the exchanges. It is a quieter corner of the Bitcoin world and a distinctly cypherpunk one. I've opened an Alby Hub, registered a Nostr account, and have been learning how to trade peer-to-peer through tools like RoboSats. The work here is exploratory rather than a position.

On My Radar

KiteAI — an emerging project I'm watching but have not yet taken a position in. I plan to write more once I've done the work to form a view.

I disclose these positions because transparency is the point of a page like this. If I write about a project, you should know whether I hold a position in it.