moonshots only | origin story of this unofficial account: https://xitter.cf/t.co/pccuYj6HRo

pi-rateship
Joined October 2024
flashbots_x retweeted
What happens when cryptography & AI builders spend 10 weeks turning research into products under real constraints? Shape Rotator Demo Day featured private agent infrastructure, self-improving workflows, sovereign data, creator tools, new interfaces + more. See what they built:
8
22
4
51
9,780
flashbots_x retweeted
Shape Rotator begins today in NYC! This is a 10-week accelerator for advanced cryptography and AI builders turning research into products. Presented by @initc3org, @flashbots_x, The Convent, and @theBBfund.
1
4
1
15
2,030
Yet another great thread on real world MEV. But you can do something about it 🔜. OAuth3: turn privacy into agency. Stay tuned @flashbots_x .
“How does MEV trading work in attention markets?” Glad you asked! Here’s a crash course in memetic frontrunning, oracle risk, and attention perps 🧵
5
4
19
7,143
On top of that there will be a virtual after party jam on Interop'ing the World with OAUTH 3 lu.ma/cnkuokby
4
1
13
4,594
flashbots_x retweeted
Andrew Miller (@flashbots_x / UIUC) @socrates1024 on "Bringing the magic of TEEs to Web2 and Web3" youtube.com/watch?v=YQwRGPtZ… 2/
3
5
11
2,917
flashbots_x retweeted
Flashbots has a surplus of great ideas for great app builders looking to disrupt web2
Replying to @ThogardPvP
hehe we are not building a wallet. there are so so many great wallet builders out there. web 2 app's dev-to-user relationship is PERMISSIONED INTEROP constrained and sanctioned by app and OS design pattern today, whereas web 3 is rooted in PERMISSIONLESS INTEROP. accounts are the portal across worlds. we are accelerating a movement that's grounded in security research that can bring together the warring factions within AA wallets, and privy's of our world, along with the zk(mpc)tls vampire attack enthusiasts together, through strict improvement in security and capabilities ;) fingers crossed, enjoy a wave of UX shock therapies across your favorite web2 apps, soon to be unleashed by the Banksy of the Web where there's no divide between 2 & 3 ;) p.s. the direction you are guessing is not wrong, but devils in the details. we may get to a similar place but here's a peek into why this matters to me: (this is part of my angry draft tweet that I never sent, after getting triggered by CT's bad vibes, and frustration seeing capable people keep getting emotionally sabotaged by tribalism that kept them from being able to work together. consider this a shower thought CTA).
4
1
24
8,299
what if those who authored Oauth 1.0 but ragequit Oauth 2.0 joins force with our Oauth3 movement? gist.github.com/nckroy/dd2d4… about time to give agency back to users - the prosumers of our future app substrates. ;) – at Denver, CO
thanks for being part of the 4am crew on @flashbots_x NYC Pi-rateship hacker village it's been a crazy fun month jailbreak web ̶2̶ Pi-rateShip sails back to SF in March to: - reorg pro-sumer relation in social apps - rearchitect data provence w/ prog privacy - reprogram the algorithms for humanity @flashbots_x X-prize: moonshots only🌕 TTEE: PUF, open source chip, verifiable fab Flashapp: user-init 0-click UX via accounts Teleport: hack attention economy substrate Dstack: secure account encumbrance OAuth3: permissionless interop the world SUAVE: geo-decentralized glue Eliza v2: putting the A back into DAOs ... fix our ruptured digital social fabric we need all the help we can get, to ship🏴‍☠️
3
4
20
4,429
thanks for being part of the 4am crew on @flashbots_x NYC Pi-rateship hacker village it's been a crazy fun month jailbreak web ̶2̶ Pi-rateShip sails back to SF in March to: - reorg pro-sumer relation in social apps - rearchitect data provence w/ prog privacy - reprogram the algorithms for humanity @flashbots_x X-prize: moonshots only🌕 TTEE: PUF, open source chip, verifiable fab Flashapp: user-init 0-click UX via accounts Teleport: hack attention economy substrate Dstack: secure account encumbrance OAuth3: permissionless interop the world SUAVE: geo-decentralized glue Eliza v2: putting the A back into DAOs ... fix our ruptured digital social fabric we need all the help we can get, to ship🏴‍☠️ – at East Village, Manhattan
Flying home after a couple months out on the road Will get back to streaming and building pretty much every day for the next few weeks Should be fun
3
3
2
82
12,459
一个小目标😉 a small target is impact >100m human users, sustainably. to help put numbers into perspective, these are the apps we can leverage today (h/t @grok deep research). YouTube: ~2.5b MAU, 1.25b DAU Instagram: ~2b MAU, 1b DAU TikTok: ~1.7b MAU, 0.9b DAU Douyin (TikTok in China): ~755m MAU SnapChat: ~800m MAU, 500m DAU Twitter/X: ~611m MAU, 250m DAU weibo: ~584m MAU, 280m DAU Pinterest: 450m MAU, 200m DAU reddit: ~850m MAU, 430m DAU Telegram: ~800m MAU, 400m DAU Discord: ~150m MAU, 80m DAU Line: ~194m MAU, 100m DAU Twitch: 140m, 70m DAU Github: ~100m MAU Figma: ~4m MAU we will be shipping multiple concurrent experiments on each platforms and across platforms, with concept artists leading different special purpose vessels of human capital formation. dm @flashbots_x if you are a siiick uiux designer, pm, reverse engineer, DevOps interested to become a TEE engineer, iOS or Android developer, chip designers, security researchers, auditors, etc.
2
4
16
2,474
flashbots_x retweeted
I’m an EVM maxi now btw
172
35
16
834
127,996
The Importance of TEE in AI Agents and the Role of Eliza Framework I believe that every AI agent should utilize TEE (Trusted Execution Environment) to ensure security and trust. In fact, I am confident that in the future, AI agents that do not integrate TEE will not survive. Why is TEE Important? - Data Protection: TEE ensures that AI agents' data remains secure from unauthorized access and modification, maintaining its confidentiality and integrity. - Verifiability & Trust: By running code in an isolated environment, TEE makes AI agent actions verifiable and reliable, increasing user trust in autonomous decisions. - Autonomy in Resource Management: TEE allows AI agents to manage assets autonomously (e.g., handling crypto transactions) without human intervention, ensuring full operational independence. What Happens Without TEE? Imagine a developer creates an AI agent without TEE integration. In such a case, the developer or other parties with access to the infrastructure could manipulate the agent’s behavior or extract sensitive data. This undermines trust in the agent and poses serious security risks. By the way, I just learned from @VirtualElena's new article that @sxysun1 and @socrates1024 (from @flashbots_x) are exploring the idea of "delegating" accounts to AI agents. They are experimenting with TEEs, creating: - AI agents that autonomously manage Twitter accounts without human intervention. - Smart contracts for accounts, where AI agents must strictly follow predefined conditions. - "Satan" – an AI that allows users to make deals, where breaking the conditions results in penalties. Key Idea: AI can achieve true agency if its behavior is fully governed by an environment inaccessible to external interference (such as TEE). The Role of ElizaOS Since I consider Eliza to be the leading framework, capable of laying the foundation for a vast number of new AI agents with the upcoming launch of its AI launchpad, I found it appropriate to specifically mention this framework. In collaboration with @PhalaNetwork, @elizaOS integrates TEE capabilities to ensure the security and transparency of AI agents. This enables AI agents built with Eliza to operate in isolated environments, protecting their operations and data from external threats. Additionally, TEE integration allows for cryptographic verification of agent actions, reinforcing trust in their autonomy.
18
18
3
135
17,178
flashbots_x retweeted
my first take on what the stages should look like: 0. agent has attestable log in TEE 1. entire agent flow (except inference) running inside of a tee 2. agent has an encumbered twitter/discord/telegram account 3. agent has an encumbered wallet 4. agent has API keys encumbered (e.g., twitter oauth dev account) 5. agent has inference inside of a TEE 6. agent has an open-source prompt 7. agent has an open-source codebase 8. agent code update time & requirements (is it a multi-sig, is it a hard time limit like can only update once per 7 days) 9. has remote attestation or not (and if yes, which tee platform, is it at a cloud?)
we need an l2beat for ai agents one that includes info on their provability, audit logs, backroom terminals, inbound data sources agents can have their own stage1 stage2 definitions
12
11
2
86
14,955
💯agentic workflow is tomorrow's app substrate how to leapfrog orderflow chokepoints, social media attention sinks? will be hacking on web2<>web3 x app<>infra at @flashbots_x NYC Pi-rateship hackerhouse tee.space in January @sxymoonsama will be judging ;)
I genuinely believe that AI agents will be the new interface layer for interacting with the internet. Communication through natural language (text + speech) is much simpler and more intuitive than following existing internet protocols. Express the action you want to do -> agent parses the request and executes it in the most optimal way. "Book the cheapest flights to London for January 14-21st with no layovers and no middle seats". AI agent searches for all possible flights, finds the cheapest ones, fills out your personal info, selects the seat, and books the tickets. Iron Man's Jarvis system will become a reality and the default norm for interaction one day.
1
4
18
5,427
flashbots_x retweeted
great stuff happening in the deAI - TEE space. Great work by @NousResearch @flashbots_x @PhalaNetwork for xitter.cf/tee_hee_he as well. Autonomous agents are about to get serious..
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
2
5
2,221
In the last few months Flashbots and the TEE/acc ecosystem have upgraded Ethereum L1, Ethereum L2, and the entire fabric of web3 social and block space auctions. It may have been easy to miss in the torrent of information, but don't worry, I'll recap it all for you below⚡️🧵
10
35
8
204
38,354
Replying to @alpeh_v
Another use case for TEE's @flashbots_x
1
1
1,105
This awards thread to wrap up the Teleport & @flashbots_x hacker house is also a concept demonstration of "confidential AI-intermediated peer review," which we are both accelerating and using for self-experiment!! (eating our own catfood). Hope for the future of coordination
This Post is from a suspended account. Learn more
1
3
16
2,623
A while back @PanteraCapital published a proof of concept using @flashbots_x' SUAVE suift.panteracapital.com/ - it describes a quite interesting credible auction mechanism where the ticket issuing smart contract would refund the difference between the auction price and the “Face value price” to everyone who ends up attending the show (so the tickets would need to be on chain for this to work). Since brokers have no intent on actually attending the event, they won't get refunded if price gauged inventory is not sold. This mechanism actually effectively influences the broker/fan utility curve in the benefit of the fans. Absolutely agree that crypto has a role to play in ticketing/events, that is why we are here.
1
1
608
flashbots_x retweeted
TEEs reinforce the integrity of @worldnetwork's Proof of Personhood in not just 1, but 3 ways. ​ Here are my key takeaways from @dcbuilder's whiteboard session during TEE Salon: _______________________________________ ​ The premise for World ID (and TEEs) ​ @worldnetwork wants to build the largest financial network for unique humans. ​ In the current age of AI, it’s becoming almost impossible to distinguish humans from AI-generated content. ​​ That’s why Proof of Personhood is so important: ​ It proves that we're unique humans which can be attested cryptographically at a global scale. ​ And at 3 stages of this process: ​ World utilises Trusted Execution Environments (TEEs) to reinforce integrity in the entire system. ​ TEEs are isolated environments with these 2 key properties: ​ ❍ Execution integrity (code runs as expected) ❍ Data confidentiality (data is not leaked) ​ But anyone can claim they're running a TEE, so we use a process called remote attestation. ​ This verifies that the TEE is indeed what it claims to be. ​ Here are the 3 ways that World uses TEEs: _______________________________________ ​ #1 Orb TEE ​ As @recmo shared by @recmo during his @EFDevcon talk: ​ Verifying the uniqueness of a human requires a trusted sensor. ​ And in World's case, this trusted sensor is the Orb. ​ Orbs are registered onchain so anyone can publicly verify that they're interacting with an authentic Orb. ​ Before they start with the World ID verification process. Each Orb has a TEE that runs integrity checks on the entire biometric pipeline. ​ Biometric data is processed in a secure enclave, preventing anyone from tampering with the pipeline. ​ Once the TEE signs on the iris image: ​ We trust that the biometric is valid and this is split into secret shares to prevent data exposure. ​ These encrypted shares are distributed to trusted nodes within World's Anonymized Multi-Party Computation (AMPC) system. ​ But there's a huge trust assumption that these nodes don't collude: ​ _______________________________________ ​ #2 MPC-TEE ​ World's AMPC system requires an honest majority to function correctly: ​ The protocol has liveness if at least 2/3 nodes are acting honestly. ​ But there’s still a risk of collusion between nodes: ​ They could choose to act maliciously which compromises the integrity of the entire system. ​ World chooses credibly neutral participants (like universities) as node operators to minimise this trust assumption. ​ But TEEs provide an additional level of integrity: ​ Running the AMPC node in a TEE ensures that the node does the thing it claims to be doing. ​ Since World’s AMPC system runs on NVIDIA’s H100 GPU: ​ Each GPU has a TEE attesting that the computations are running correctly and as expected. Once the user's biometric data is processed and cross-verified to be globally unique: ​ The user will be issued an anonymous credential (World ID) on their phone. ​ But the Orb has to check that it's interacting with an authentic device first before issuing this credential: ​ _______________________________________ ​ #3 App (Client) TEE ​ All of our phones have a TEE: ​ Features like App Attest (iOS) and Play Integrity API (Android) verify the authenticity of the app and device. These require the device's secure enclave (TEE) to sign and attest that the device is running trusted software. ​ TEEs within the app client and the Orb communicate with each other in these ways: ​ ❍ The phone proves to the Orb that it’s a real phone running an authentic OS ❍ The Orb proves to the phone that it runs a correct implementation of the Orb software ​ World eventually aims to decentralise the entire model so that Orbs can be manufactured by anyone. ​ Even though we can’t trust the parties involved (app and Orb) since it’s all decentralised: ​ We can trust that they’re running authentic software through TEEs. And this maintains the integrity of the entire system. ​ The main app for now is World App, but there could be other app clients in the future. ​ _______________________________________ ​ Here's a summary of the 3 ways World uses TEEs for Proof of Personhood: ​ ❍ The Orb TEE attests that the biometric data was collected with a valid process ❍ TEEs in the NVIDIA H100 GPU attest that AMPC nodes are running correctly ❍ TEEs in the client and Orb communicate with each other to verify they're interacting with authentic Orbs and clients ​ And at @AutomataNetwork, we're working with @worldcoinfnd to secure AMPC nodes with TEEs. ​ I explained how TEEs fit into the AMPC system in this post: ​ xitter.cf/nuhgid/status/18473074… ​ _______________________________________ ​ Other resources that may interest you ​ ❍ How I'd explain TEEs to my former self ​ xitter.cf/nuhgid/status/18449479… ​ ❍ Open sourcing and auditing the Orb with @reldev, @philzip, @DanGirsh, and Ryan Butler ​ youtube.com/watch?v=50L_1P7y… ​ Thanks for reading! ​ If you want more like this: ​ ​ 1. Follow me @nuhgid 2. Like, RT, and share this post with anyone eager to learn more about cryptographic primitives in Web3
TEEs don’t have to be complicated. ​ It took me 11 months at Automata to finally get the hang of it, but you need less than 10 minutes to understand it fully. ​ Here’s how I’d explain TEEs to my former self: ​ _______________________________________ ​ The black box analogy ​ After a plane crash, the first thing that investigators will look for is a flight recorder, AKA a black box. ​ (Which in fact, is not actually black but orange in colour instead) These flight recorders have some similarities to a Trusted Execution Environment (TEE): ​ #1 Isolation ​ The flight recorder is isolated from the main aircraft systems so it continues to record flight logs even when these systems are down. ​ A TEE is also isolated from the main operating system. Even in a compromise, the operations within a TEE can still run and remain unaffected. ​ #2 Execution integrity ​ The flight recorder is meant to resist harsh conditions. So the flight data remains intact for investigators to analyse. I found this picture of a flight recorder from the 2016 crash of West Air Sweden Flight 294. All data was collected, even though the aircraft was badly damaged. ​ A flight recorder guarantees that the data is collected accurately. ​ A TEE also enforces execution integrity, where we know the computation will be completed exactly as intended. ​ The same input will always give the same output, no matter the condition. But a flight recorder is not a fully accurate analogy of a TEE due to this difference: _______________________________________ Keeping data truly secure​ After an incident, the data in a flight recorder is meant to be retrieved and analysed openly by investigators. Data is only protected from these factors: ❍ Physical damage ❍ Unauthorised access (only those with the correct credentials can access it) But a TEE (if functioning properly) is different: It prevents anyone from accessing the data. Not even the hardware manufacturer can access it. In this way, whatever goes into a TEE becomes the world's best-kept secret: ​ No one will ever know what data is processed in the TEE. ​ Flight recorders also lack this feature that helps us trust computation from a TEE: _______________________________________ ​ The secret ingredient to trusting a TEE ​ I like @sbellem's explanation of what the main goal of any TEE is: ​ We can send computations to a remote machine and be confident they are executed as expected without trusting the machine operator. ​ But here's the main issue: ​ Any machine can claim they are operating a TEE. ​ So TEEs have a feature called remote attestation. ​ This enables any party to verify the authenticity of a TEE. (Even those who are not in physical contact with the machine) ​ Once we confirm that the TEE is what it claims to be: ​ We have greater trust that: ​ ❍ Sensitive data won't be leaked (Confidentiality) ❍ Computations will executed as expected (Integrity) ​ But how do we verify this authenticity? _______________________________________ ​ The hardware root of trust ​ A private key securely stored in the hardware signs an attestation certificate during the remote attestation process. ​ Unlike software which can be more easily attacked or manipulated: ​ A hardware-based root of trust offers a stronger level of protection that is difficult to compromise. Since the chip is embedded during the manufacturing process, it cannot be modified as easily as software. ​ However, there are still some concerns that current closed-sourced TEEs are not 100% secure. ​ @Poetic_Tech hosted an X Spaces on this topic, and I shared my takeaways below: ​ xitter.cf/nuhgid/status/18443457… ​ _______________________________________ ​ Putting it all together ​ Here's how I understand TEEs after combining these concepts: ​ A TEE runs code in an isolated environment that is secured by the hardware root of trust. ​ Due to its isolation, it continues to operate even when the main system is down. Just like a flight recorder. ​ A TEE is tamper-proof, and no one can modify or view the code/data being processed. ​ This leads to the 2 key properties of a TEE: ​ ❍ Execution integrity (code runs as expected) ❍ Data confidentiality (data is not leaked) We need to verify that a TEE is what it claims to be before we can trust it with code or sensitive data. So most TEEs come with a remote attestation feature so anyone can verify its authenticity. Even those who are not in physical contact with the machine. _______________________________________ ​ Still confused about TEEs? ​ My team and I from @AutomataNetwork built an ELI5 site that explains TEE with the story of Enclavus: ​ A barren wasteland with scarce resources. ​ Check it out below: ​ eli5.ata.network/ ​ Thanks for reading! ​ If you want more like this: ​ ​ 1. Follow me @nuhgid 2. Like, RT, and share this post with anyone eager to learn more about TEEs in Web3
2
8
1
55
5,680