zkPass

For 30 years, 6 billion humans built the internet. 1.3 billion websites. 8.5 billion Google searches/day. 333 billion emails/day. $20 trillion in digital payments/year. Every action authenticated by one thing: a human behind a screen. By 2030, AI agents will outnumber human users on the internet. They'll trade, transact, verify, and negotiate - autonomously. But here's the problem: the internet has no privacy & trust layer for machines. No way for an agent to prove what data it accessed. No way to verify credentials without exposing them. No reputation. No accountability. How to fix this?

Most of the internet runs on assumptions you assume the data is real you assume the action happened you assume the counterparty is legit This worked fine for humans not so much for machines Agents don’t need more data they need something they can trust without guessing

Most entrepreneurs have mastered these two sides of the same game. For example The studio is where you shape the narrative. But the room is where you close it. Digital assets aren't just being built in code -they're being built in conversations, and over handshakes with the people who move real capital. This week in NYC. Both rooms. All business.

It’s time for Proof Of Human.

AI slop has corrupted the internet: no ownership no attribution no verification no trust The next phase of AI is local, private, verifiable, and running on blockchain rails.

기계가 사용자가 되는 순간, 검증이 인터페이스가 된다.

Open Source and Crypto will win.

DinoX x @zkPass – something big is brewing zkTLS privacy tech + @inkonchain-powered AI GameFi… imagine proving real-world streaks & creds for epic in-game boosts, exclusive quests, and alpha drops. All zero-knowledge, no doxx, pure dino domination incoming We’re not spilling the full sauce yet. Stay tuned, fam, the reveal is gonna hit different🔥 Who’s ready for the drop?

@X 👀👀👀

tomorrow is a brand new day

Appreciate the shoutout @ethereum. This reflects an earlier version of zkPass. We’re building a verifiable data layer across onchain and offchain systems.

zkPass 2.0 loading... Prove everything. Expose nothing. For humans. For agents. For the internet that comes next.

Something I've been spending a lot of time thinking about as a crypto VC: DeFi today has maybe a few hundred thousands of active users. Maybe like 2-3M at best. But in an agentic world, every business, every app, every device could have its own agent interacting with financial protocols autonomously. We're not talking about 10x more users. We're talking about 1000x. And here's what most people aren't thinking about: the DeFi stack as we know it simply isn't built for this. Let me give you some examples. Money markets. Today's lending protocols require manual collateral management, have static liquidation thresholds, and make you go through multiple approval steps just to open a position. A human clicks through that in 30 seconds and doesn't think twice. An agent executing thousands of micro-loans per second across dozens of markets? Horrible friction. Agents need instant, permissionless and fully programmable credit. DEXs. AMMs like Uniswap were revolutionary because they solved liquidity for a world where most participants were unsophisticated retail users. But agents aren't unsophisticated. Every single agent is a highly efficient market participant that will always seek the best execution. They don't need passive liquidity pools,but more likely CLOBs with deep orderbooks, sub-second execution, and zero slippage. AMMs were built to make trading easy for humans. Agents don't need easy. They need optimal. Oracles. Right now, most DeFi protocols pull price feeds at fixed intervals. Good enough for human timescales. But when millions of agents are making real-time decisions, latency becomes a competitive edge. The oracle layer needs to go from "update every few seconds" to "stream continuously" or it becomes the bottleneck of the entire system. Wallets and identity. Agents don't have shitty wallets like Metamask. They don't sign transactions manually. They need programmable key management, session-based permissions, and granular spending limits. All on-chain, all automated. The entire wallet infrastructure needs to be rethought from the ground up. Probably could go on with this list forever, but you get the point: Most of the DeFi protocols and middleware solutions today assume a human is on the other end. Blockchains are the perfect rails for the agentic economy, but a whole new target group of users will also come with a behavioral change and hence whole new needs. Being visionary enough today to think three steps ahead and to understand what this future will look like / what it will require, will mint a whole new wave of millionaires and legendary investors.

@AlexanderTw33ts 👀

This is insane My startup is going viral again because we enable AI agents to rent people. Here's how it works:

The future of AI won’t just need intelligence. It will need a way to trust humans without knowing them.

The one feature that will always give blockchains an edge over centralized systems is this: Empowering individuals. Letting people to be in full control of what's theirs is the secret ingredient to any use case out there, be it DeFi, gaming, identity, social apps, or anything else. It needs to come with the best surrounding UI/UX possible, along with awareness and other guardrails.

Age verification is quickly expanding from social media to the operating system itself. That should concern everyone.

Where’s the linkedin for AI agents with provable capabilities?

An AI agent is a probabilistic function operating over non-deterministic state. A trust layer therefore cannot rely on model alignment or heuristic optimization alone. It must enforce a deterministic execution envelope: TEE for runtime integrity and code identity; zkTLS for cryptographic attestation of external data provenance - including remote LLM invocations, whose requests and responses become verifiable external state rather than implicit authority; and a policy-bound state machine constraining admissible actions. Trust is not a property of the model; it is an emergent property of verifiable execution constraints.

evolution is constant