Founder of Rivellum - Intent-native Layer 1 with zero MEV, no contract custody, horizontal scalability, post-quantum, and compliant ZK privacy.
Massachusetts, USA
Joined June 2019
- Tweets3K
- Following592
- Followers305
- Likes2.8K
Ben Abbott retweeted
AI agents are already executing financial transactions autonomously. This isn't a future scenario.
And the infrastructure they're running on wasn't built for them.
Give an agent a standard wallet on a standard blockchain and you've given it unlimited financial authority within its key. One compromised key. One bad prompt injection. One exploited integration. The agent can spend everything, call anything, trigger any transaction the wallet allows.
Agents don't need unlimited authority. They need bounded authority.
What autonomous AI agents actually require from blockchain infrastructure:
Protocol-level execution bounds - the agent can only spend within defined limits, call approved services, and move value under specific conditions. Not enforced by the application. Enforced by the protocol itself.
Predictable fees - an agent managing workflows across thousands of transactions cannot function if fee estimation is unreliable. It needs to budget deterministically.
Privacy by default - an agent's pending transactions should not be visible before execution. Exposed strategy is exploitable strategy.
Guaranteed outcomes - the agent needs to know whether its action succeeded or failed with certainty, not probability.
No contract custody risk - agent funds should not sit inside application logic where a bug becomes a drain vector.
Rivellum is being designed to satisfy all five of these requirements at the protocol level. Not as application-layer tooling bolted on top of infrastructure that wasn't designed for it. As architecture.
The machine economy needs infrastructure built for machines.
We're building it.
rivellum.network
Every trust boundary covered from genesis:
Account keys. Validator identities. Consensus messages. Intent authorization. Settlement commitments.
Top chains are announcing upgrades, yet most will remain exposed somewhere vital to the network.
Rivellum was built prepared for the future.
In August 2024, NIST finalized its post-quantum cryptography standards after an eight-year effort. Their guidance to system administrators was explicit: begin transitioning immediately.
Most blockchain infrastructure hasn't moved.
The threat isn't just future quantum computers breaking future transactions. It's harvest now, decrypt later.
Adversaries are storing encrypted blockchain data today — account keys, validator credentials, transaction signatures — planning to decrypt it when quantum computing matures. Every account created under classical cryptographic assumptions is a future liability. Every validator key. Every long-lived credential.
Migration sounds simple. It isn't. It's not swapping one algorithm for another. It's every account key, every wallet authorization, every validator identity, every consensus message, every proof verification, every bridge attestation. If any of those surfaces launches under classical assumptions, the migration problem exists from day one regardless of what you do later.
Rivellum is being designed with Dilithium3 — NIST's highest security tier digital signature standard — covering every critical trust boundary in the protocol from the start.
Account keys. Validator identities. Consensus messages. Intent authorization. Settlement commitments. All quantum-resistant by design. Not as a planned upgrade. As the foundation.
Every other chain will eventually have to announce a migration. We're building Rivellum so it never has to.
rivellum.network
This is one of the core reasons I started building Rivellum.
MEV isn’t just something to mitigate or share - it’s a structural problem that shouldn’t exist in the first place. Encrypted intents + blind ordering removes it at the root.
Estimates put total MEV extraction across major chains at over $2B in 2025 alone.
Not from hacks. Not from exploits. Taken legally, every day, through transaction ordering — from ordinary users who never knew it happened.
Here's the mechanic:
You submit a swap. Before it executes, a bot sees it in the public mempool. It front-runs you — buying first, driving up your price, selling immediately after your transaction confirms. You paid more. The bot profited. You had no idea.
Most chains try to mitigate this. Priority auctions. MEV-sharing protocols. Proposer-builder separation. These don't eliminate the problem. They redistribute who profits from it.
Rivellum is being designed differently.
Intents are submitted encrypted. Ordering is assigned blind — no participant sees transaction content before the ordering decision is made. By the time content is visible, ordering is already fixed by cryptographic seal.
You cannot front-run what you cannot see.
This isn't a configuration. It's the architecture. MEV elimination isn't a feature we added. It's a consequence of how the protocol is designed from the ground up.
Currently in pre-funding development. Building this in public.
rivellum.network
Attn Rustians!
Rivellum is looking for a CTO / technical co-founder.
Live testnet, parallel lanes, Mist, UVL, post-quantum, real benchmarks.
If you're a strong Rust / protocol engineer and this resonates, DM me.
US preferred but open to remote.
The evolution has arrived. Excited to see what builders and researchers think 👇
The evolution has arrived.
Rivellum, the intent-centric Layer 1 for guaranteed outcomes, is now live in very very early beta testnet (temporarily for limited testing).
The protocol replaces transactions with intents and constraints. It resolves them deterministically — or rejects them cleanly before execution.
No mempool. No ordering games. No contract custody.
📘 Whitepaper: rivellum.network/whitepaper.…
Testnet open for early feedback → rivellum.network
Follow @rivellum and turn notifications on.
Ben Abbott retweeted
cheers to the real ones 💫
global @Aptos Community team and @AptosCollective members repping at AE!
thanks for being the crew that keeps the ecosystem connected and growing - welcoming new folks and helping with education and onboarding 🫡
📷: Aptos Labs + Foundation @AveryChing @ASHAWONN @Greg_Nazario @NareshKx @NickGCat
Community Team: @0xayushya @angelataptos @Gemswizzard @realjaysneezy @stan_ngx @hawaiianft
Collectives + extended community: @0x_Baba @kantianum @NathMAN90 @hiAtreyu_ @ApersonLyeque
@AptosGato @juanmsobral @Cryptofinomic @setofsevens
Ben Abbott retweeted
We are building the best infra for trading.
The big alpha I am keeping for Aptos Experience.
Make sure to be there. Oct 15-16 NYC.
When it comes to alpha, always look to the builders.
Our Head of Research @sashaspiegelman breaks down the latest tech driving @Aptos forward—and drops a glimpse of what’s next at Aptos Experience 👀
Couldnt have said it better myself. Propa to Aptos for looking for ways to reward creators and thibking outside the box, but this was not the answer. Excited to see what they plan with season 2. Sure hope the bullishness and creators continue to create the awesome content ive grown accustomed to for Aptos!
I'm never glad that such campaigns end (even if temporarily), but I'm particularly glad with this particular one because, even though @Aptos wanted creators to win, the leaderboard was being farmed by groups and literal cabals. Thus, I am happy to see the end of Season 1.
Also, glad that it was the Aptos Kaito leaderboard and discussions that led to me and @kantianum starting to chat with each other, and now we are pushing the content distribution arc to the max on Aptos with @minds_DAO!
Check for the eligibility for the last month's leaderboard, and stay tuned for what's next. And if you are a creator, follow Minds DAO and join the Discord, because the next couple of months are going to be filled with some creative fun alongside the content contests!
Ben Abbott retweeted
Breaking: 100ms blocks to ~60ms blocks (40% reduction) coming to @aptos with optimistic proposals in AIP-131.
100ms blocks were already industry leading. But not enough for the exchange use cases on Aptos. The global trading engine needs more speed to get faster price updates, adjust positions, and avoid stale trades. The era of decentralized trading is just starting and radical vertically designed infra will bring forth new and expansive product experiences.
Ben Abbott retweeted
Your weekly dose of organized chaos! xitter.cf/i/spaces/1yNGabEaYPWJj
