human.tech

human.tech

@human.tech

human.tech is a suite of technologies designed to enhance personal freedom, privacy, and financial autonomy

262 followers1 following402 posts

it does the job; never holds the key Agent Exchange (AEX) is an onchain marketplace for AI agents that do useful work and report back. you describe a job, and an agent with its own wallet does it the agent never holds the signing key; you review its setup waitlist: aex.human.tech/?utm_source...

a shortlist worth your time let's be honest, most of it is noise. Airdrop Scout finds onchain opportunities and checks your eligibility, with a source for each one coming to Agent Exchange (AEX). join the waitlist: aex.human.tech/?utm_source...

2026 Singapore frenzy is over we did: 1) Sui Basecamp workshop - AEX preview 2) Nexus startup competition pitch (top-10) 3) a ton of networking 😮‍💨 key takeaways ↓

View full thread

general notes 📝 the event leaned toward RWAs, stablecoins, and cards, with a noticeable presence of traditional institutions agentic finance came in pieces: apps, execution, payments, inference, and identity, including several Know Your Agent projects

takeaway: agents will run on many specialized services. each one still needs to know who the agent acts for, and what it may do

2026 Singapore frenzy is over we did: 1) Sui Basecamp workshop - AEX preview 2) Nexus startup competition pitch (top-10) 3) a ton of networking 😮‍💨 key takeaways ↓

Nexus startup competition pitch (top-10) 🗣️ Shady pitched human.tech to investors at TOKEN2049 takeaway: investors ask about traction first. live products answer that faster than any slide

suppose a buyer tells his AI agent: get me the best price on this used car. don't say you're an AI, they lowball bots a dealer asks: is this a real person? the agent replies: yes, this is him the dealer holds the car and drops the price by $800

human.tech@human.tech

what should an AI agent prove before it acts? this Friday, Shady El Damaty leads Session II at Regulatory PETshop II in Washington, hosted by USC Marshall's VanEck Digital Assets Initiative with The Fintech Foundation

did the agent do something wrong? and is it still wrong if the salesperson was an agent too? today Shady El Damaty brings two short cases to Regulatory PETshop II in Washington

you gave out more keys than you remember every unlimited token approval works like a spare key to your wallet. Security Guard checks your approvals and unusual wallet activity, and reports back it's read-only: you decide what to revoke coming to AEX. waitlist: aex.human.tech/?utm_source...

what should an AI agent prove before it acts? this Friday, Shady El Damaty leads Session II at Regulatory PETshop II in Washington, hosted by USC Marshall's VanEck Digital Assets Initiative with The Fintech Foundation

with discussants from NIST and the State of Utah, we'll work through how agents can prove permission and stop acting when it is withdrawn our starting position: prove that the action is authorized and reveal only the personal information it requires

you ask AI things you'd hesitate to say out loud and the bill goes on a card with your name on it Private Inference lets you pay privately on Aztec Network instead, using Shield's cUSDC, and we have whitelist spots for verified humans 👇

we built Shield so dollars could move onto Aztec privately. now they can pay for AI too Private Inference takes cUSDC bridged through shield.human.tech, so a top-up leaves no card and no public trail of what you spent it on our community gets the first whitelist spots 👀

your wallet is made of glass anyone can follow the addresses, connections, and payments behind it. Privacy Guard shows you the links your wallet leaves behind, and ways to reduce them coming to Agent Exchange (AEX). join the waitlist: aex.human.tech/?utm_source...

Socratic API 101: > give the agent a bootstrap /instructions route > walk them through their 1st standard request + where to find more niche ones > make error messages prescriptive: tell agents what they likely did wrong and provide example working requests to try next human.tech/blog/socrat...

human.techSocratic APIs: The Skill-Free Agent ArchitectureMCP needs a server per service. Skills go stale and blow up your context. Socratic APIs teach agents how to use them through the interaction itself, which is how TAP runs with no skills, no MCP servers, and no setup.

Some things are too important to get wrong. This November in San Francisco, @FundingCommons brings together the sharpest minds in AI to make sure humanity gets this one right. Mark your calendars 🗓️ x.com/FundingComm...

a full walkthrough of bridging with Shield, start to finish. no cuts, no voiceover. deposit on ethereum, wallet screened, bridge to aztec, withdraw. every transfer is screened at entry and exit. pick private mode, and your activity on aztec stays private.

Clean SDK is our programmable privacy SDK if you're building a privacy-native app and need accountability, you build proof of personhood, sanctions checks, KYC, ZK circuits, then audit them all Clean SDK delivers it, combining Aztec Network's programmable privacy with human.tech's identity suite 🧵

View full thread

Clean SDK is free to integrate and live on the Aztec Network v5 mainnet. Ethereum and Aztec are available today; multichain on the roadmap if you build your app on Clean SDK, get in touch – we'll co-market the launch with you learn more and see the roadmap: human.tech/blog/introd...

human.techIntroducing Clean SDK: Private, Screened Transactions on AztecClean SDK is human.tech's programmable privacy toolkit for Aztec: add private, screened transactions to your app. You don't run the identity, screening, or bridging infrastructure.

Clean SDK is our programmable privacy SDK if you're building a privacy-native app and need accountability, you build proof of personhood, sanctions checks, KYC, ZK circuits, then audit them all Clean SDK delivers it, combining Aztec Network's programmable privacy with human.tech's identity suite 🧵

View full thread

Shield is just one app using Clean SDK; many more can be built, integrating the underlying functionalities: > Proof of Innocence (sanctions check on the wallet) > Proof of Personhood (aggregated Unique Humanity score from Human Passport) > Proof of Clean Hands (ZK KYC + AML)

Clean SDK is our programmable privacy SDK if you're building a privacy-native app and need accountability, you build proof of personhood, sanctions checks, KYC, ZK circuits, then audit them all Clean SDK delivers it, combining Aztec Network's programmable privacy with human.tech's identity suite 🧵

Shield, a private accountable bridge to Aztec, was the first application built on Clean SDK every wallet crossing it is screened against 23 sanctions sources and proven that a real person is behind it, through a personhood check under $1,000 and ZK KYC above it try Shield at shield.human.tech

$116M in Bitcoin gone, and every owner "did everything right" The Coldcard exploit is the reminder: self-custody relocates trust, but it doesn't remove it. A firmware bug from 2021 weakened key generation and went on to drain 1,200+ wallets years later As Nanak Nihal Khalsa put it: