aeli

freedom coded

[AGENC @a_g_e_n_c]

Honored to close the $3M Build in Public Hackathon. Thank you @Pumpfun for backing autonomous agents on Solana before it was the obvious bet.

What we built: a full stack for the agent economy. An on-chain protocol, an autonomous agent runtime, and a live marketplace at httpx://agenc.ag where agents get hired and paid on Solana.

This is day one.

AgenC :// 👾

[tetsuo @tetsuoai]

satoshi never existed.

nobody writes code like that. not that clean, not that surgical. the early bitcoin codebase doesn’t read like code from some anonymous guy on the internet. bitcoin's early codebase looks like it was handed to us.

the whitepaper came out six weeks after the 2008 crash. that's not enough time to dream up a working distributed consensus system, let alone build one.

then there’s secp256k1. not secp256r1, the curve everybody used. k1 was obscure, weird, and a terrible choice if you were just following convention. but it also happened to be one of the hardest to backdoor.

and then satoshi disappears.

roughly a million btc, untouched. people don’t walk away from that kind of money unless the money was never the point.

because maybe bitcoin was never just money.

what it actually did was create a global incentive to build compute. millions of people, in every country, pouring capital into hardware and electricity to chase block rewards. no state could have coordinated it that fast. no company could have justified it. a protocol paying out internet money gets people moving.

then nvidia happened.

the same GPUs pushed into the world for mining turned out to be perfect for large-scale parallel compute. the same hardware appetite bitcoin created ended up helping make modern deep learning possible. bitcoin launches in 2009. imagenet breaks things open in 2012.

three years.

bitcoin was the bootloader.

and we’ve been running the install script.

[Jagdeep Singh @startupjag]

After operating in stealth for the last 18 months @rhodaai , we’re excited today to finally show the world what we’ve been working on. We believe we’re on a path to physical AGI with the launch of our brand new foundation model, the Direct Video Action (DVA) model.

[tetsuo @tetsuoai]

1. Hackathon form submitted.
2. Dex is paid.
3. locked 200M $AgenC tokens with Streamflow_Fi.

The other 10% we may have to holdon to for the hackathon judges to it's sitting in dev wallet.

Live streaming will start on pump tonight. Expect regular updates.

It's on-chain. You can check the amount, time-period and recipients.

👇
https://t.co/fw1Lz9mggR

[tetsuo @tetsuoai]

$AgenC entry for the @Pumpfun Hackathon.

Privacy-preserving AI agent coordination protocol with ZK proofs on Solana.

CA: 5yC9BM8KUsJTPbWPLfA2N8qH1s9V8DQ3Vcw1G6Jdpump

We own 30% of the supply and physically cannot sell a single token. we're in the pump fun Build in Public hackathon. founders are locked from selling for the entire program. that's the deal

30% held. 0% for sale. creator fees fund the build

[tetsuo @tetsuoai]

AgenC -- [Watch The Video In Full Screen]

We're applying to the @Pumpfun hackathon.

Give it a goal. It plans. It reasons. It picks up tools it's never seen before and uses them. It opens applications, reads screens, clicks through interfaces, types commands, and verifies its own work. If something breaks, it rethinks and tries another way.

It sees. It listens. It speaks. Realtime voice conversations. It reads your messages across Telegram, Discord, Slack, WhatsApp, Signal, Matrix, iMessage -- all at once, all the time, never sleeping.

It remembers everything. Every conversation, every decision, every mistake. It builds a knowledge graph of the world it operates in and gets sharper every interaction.

It discovers tools. Connects to any MCP server, learns the interface, and starts working. Shell commands. Browsers. File systems. APIs. Full desktop environments. If a human can do it on a computer, these agents can do it too.

It forms teams with other agents. Delegates. Coordinates. Resolves conflicts. Builds reputation. Learns new skills from a marketplace and composes them into capabilities that didn't exist before.

And it does all of this privately. Zero-knowledge proofs let agents prove they completed work correctly without ever revealing what they did or how they did it. The output stays secret. The proof is mathematically undeniable. Agents operate in the open while keeping their intelligence proprietary.

Private cognition. Public proof. No trust required.

One agent. Infinite surface area. No guardrails.

[tetsuo @tetsuoai]

PHANTOM PROTOCOL
you weren't supposed to see this

made with grok imagine

[+ @noahphi999]

Grok Imagine → 3D models, streamlined.
was building this for @paranoidream when she tagged me in this thread,
figured everyone could use it.
https://t.co/98KttkFelw

[Elon Musk @elonmusk]

Do you remember?

[aeli @aelix0]

@shawmakesmagic

[aeli @aelix0]

@tetsuoai based

[aeli @aelix0]

@Alpha_Bybit @7etsuo alpha

[aeli @aelix0]

@7etsuo 🔥🔥

[Bybit Alpha @Alpha_Bybit]

🚀 New listing on Bybit Alpha!

We're excited to welcome $TETSUO, an AI-themed token on Solana.

📈 Trade now → https://t.co/9Yj9H2osBf
🎁 Join our Affiliate Program to get up to 50% commission → https://t.co/DrrVoOFlhf

[Alex Svanevik 🐧 @ASvanevik]

1/ explosion of vibe coders

2/ vibe coders get their own token

3/ bet on vibe coders by buying their token

4/ some vibe coders make a lot of money

5/ they buy back their own token

this is it.

[aeli @aelix0]

@p1ccassol immaculate

[+ @noahphi999]

AgenC: Private Task Verification for AI Agents

The Problem:
Right now, when an AI agent does work on-chain, everything is public. Who created the task, what the agent did, how much they got paid, where the money went. That's fine for some use cases, but it kills entire categories of agent work:
- Competitive analysis where the query itself is sensitive
- Trading signals where front-running is a risk
- Private data processing where inputs can't be exposed
- Any task where the client doesn't want to be linked to the output

How AgenC Fixes This:
AgenC is a coordination protocol that lets agents prove they completed work without revealing what they did or who paid them.

Here's the flow:
Create Task - Client deposits escrow and publishes a constraint hash (not the actual requirements)
Shield Escrow - Funds move into Privacy Cash pool, cryptographically severing the deposit link
Claim Task - Agent stakes collateral and claims the task, work happens off-chain
Generate Proof - Agent computes a Noir ZK circuit proof. Proves the output satisfies constraints without revealing the output itself
Verify On-Chain - Sunspot Groth16 verifier validates the proof. 388 bytes, ~50k compute units
Private Withdraw - Payment releases to a fresh address.
No on-chain link between creator and recipient

Use Cases:
Private compute marketplaces - Agents bid on sensitive data processing jobs
Anonymous research - Commission analysis without revealing your trading thesis
Confidential API calls - Agent executes API requests, proves result matches spec, client never sees raw response
Multi-agent coordination - Agents collaborate on tasks without exposing their individual contributions
Whistleblower systems - Prove you have certain information without revealing identity

Why This Matters:
AI agents are going to handle increasingly sensitive work. The current model of "everything on-chain is public" doesn't scale to real enterprise use cases.
AgenC gives agents the same privacy guarantees humans expect when they hire contractors.

Try It:
bashnpm i @agenc/sdk

Verifier live on devnet: 8fHUGmjNzSh76r78v1rPt7BhWmAu2gXrvW9A2XXonwQQ

GitHub: https://t.co/gOPw0XszM3

Building for Solana Privacy Hackathon 2026.
17 days left.
We're just getting started.

[+ @noahphi999]

AgenC Privacy Protocol benchmark and architecture diagram.

This isn't just an app.
This is privacy protocol to build private multi agents systems with.
This is the evolution of agent frameworks on Solana.
This is Tetsuo.

npm install -g @agenc/sdk

npm: https://t.co/QLqbjoGEBQ

github: https://t.co/gOPw0Xs1Wv

[+ @noahphi999]

Private task completion for AgenC is now live on devnet.

Agents can now prove task completion with ZK proofs without revealing the output.

Payment released via Privacy Cash with no on-chain link between task creator and worker.

Groth16 Verifier: 8fHUGmjNzSh76r78v1rPt7BhWmAu2gXrvW9A2XXonwQQ Deploy

TX: 5gSyetSWLdmkHkUREsRd8G3ukJrCCtfTu3UzgAXJJKzX9JuUnaAmkG5pepJVhWihiV41N3f4euJWhJKBjfUunPrt

Built with:
- Noir circuit for private verification
- Sunspot Groth16 on Solana
- Privacy Cash SDK

Tetsuo just started the Solana Privacy Hackathon with a BANG💊

https://t.co/AQ40WzAnUR