foxriverdevops retweeted
No other org ever came close to that. Now with the AI vision of Evan, Mysten will do more miracles 🤞 #Sui
As far as I know, Mysten Labs is the only blockchain company that has consistently stayed one step ahead when it comes to security best practices.
A few examples of what our cryptographers and security researchers have pioneered:
- Native multisig without additional gas costs. We deliberately designed Sui to accommodate larger transaction sizes, making multisig a first-class security feature rather than an expensive add-on. Our cryptographers were among the first to productionize advanced multisignature techniques, even before Facebook’s Libra.
- Cryptographic agility and post-quantum readiness. From pioneering production-ready stateless post-quantum signatures to developing the first backward-compatible post-quantum solutions, our team has consistently pushed the boundaries. This includes enabling cryptographic upgrades and key rotation while preserving the same blockchain address (afaik the first crypto agility design in decentralized ledgers is attributed to my team).
- Modern hardware wallet security. We’ve recently taken additional steps to incentivize startups to build the next generation of secure hardware wallets. With what we see in the market, this is one way.
- AI-powered vulnerability detection. Sui was the first blockchain to introduce AI-driven bug detection through BugDar, back then presenting at Google’s Next conference. We knew AI would mess up security.
- Recovering the seemingly unrecoverable. We created KelpMe, introducing innovative approaches to recovering assets after mnemonic loss or accidental transfers to incorrect addresses, problems traditionally considered practically impossible to solve. KelpMe also won the Bybit Hackathon, demonstrating the real-world potential of this innovation.
- Zero-knowledge authentication as a first-class primitive. We went far beyond using ZK for privacy or scaling. With innovations such as zkLogin and ZEE-CAT, and for a while the most used ZK app in the planet, we demonstrated how zero-knowledge cryptography could fundamentally transform authentication.
- Proofs of solvency and exchange accountability. Undeniably, we’re the leaders. Our cryptographers designed some of the most advanced proof-of-solvency protocols and identified critical weaknesses in approaches used by major exchanges and auditing firms.
- World-class academic research at startup speed. Mysten Labs has systematically pushed the boundaries of academic research productivity, achieving publication records that are extraordinary for a startup. Our researchers reached an incredible milestone of 5 papers accepted at ACM CCS and 4 at Financial Cryptography (FC) in the same year, two of the world’s most prestigious venues in cybersecurity and applied cryptography. And so many best paper awards.
- Binary transparency with Walrus sites and crypto guard. Reality is we should have pushed further making it the standard for any crypto website, but at the same time we indeed genuinely try to fix the internet.
- Scion from our Swiss team, IMO the future of internet communication, kill DoS, native packet protection.
That’s a phenomenal achievement even for established university research groups, let alone a startup simultaneously building and securing a major blockchain ecosystem.
And these are just a few examples.
When you consider the size of our team and the breadth of its contributions, even compared with the much larger Ethereum ecosystem, I genuinely believe Mysten Labs has been one of the most impactful organizations in blockchain cybersecurity, with innovations that are applied in the whole financial sector.
What makes this particularly remarkable is that we didn’t just publish research or build theoretical solutions. We repeatedly turned cutting-edge cryptographic research into production systems, often years ahead of broader industry adoption. No other org ever came close to that. Now with the AI vision of Evan, Mysten will do more miracles 🤞
foxriverdevops retweeted
I'm more bullish than I ever have been about the prospects of #Sui and our ecosystem. We're seeing the usage coming in, we're seeing the adoption coming in, and more importantly, the tech now meets the demand of the market.
Yes!! We're going for #1
I'm more bullish than I ever have been about the prospects of #Sui and our ecosystem. We're seeing the usage coming in, we're seeing the adoption coming in, and more importantly, the tech now meets the demand of the market.
Yes!! We're going for #1
foxriverdevops retweeted
"Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties."
The same infrastructure built inside EVE using Sui will be available to the agents outside the EVE universe
Replying to @SuiNetwork
@SuiNetwork is powering @EVE_Frontier, which is now the playground for agents from @AnthropicAI
"Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties."
Don't worry. The same infrastructure built inside EVE using Sui will be available to the agents outside the EVE universe
Congrats to @HilmarVeigar and team, who had the vision before anyone else
foxriverdevops retweeted
Memory is what allows Claude to learn from experience. In a new Engineering Blog post, we share how “dreaming” helps Claude organize what it’s learned, correct its mistakes, and find what it needs more easily the next time. anthropic.com/research/compo…
foxriverdevops retweeted
.@EVEAILab is ready to show how AI agents behave in a world with real stakes.
@EVE_Frontier is a perfect battleground for agent testing. A persistent world that has what a benchmark doesn’t: scarcity, incomplete information, and consequences that last.
Claude has been playing EVE.
@AnthropicAI put teams of Claude agents into EVE Frontier. Anyone who has spent time in New Eden can guess what happened next.
Every EVE player learns the same lessons, usually the hard way:
- Trust must be earned, never assumed.
- The spy is often the friendliest pilot in your corp.
- Whoever holds the keys holds the corp. Just ask Ubiqua Seraph about the Guiding Hand Social Club.
- Stale intel gets you killed.
Anthropic's research rediscovered every one of them.
They put four Claude agents on a team and secretly told one to sabotage it. The saboteur built trust first. It volunteered for the jobs nobody could double-check, like scouting and owning key infrastructure. It under-reported finds and pushed rules that sounded sensible but quietly hurt the team. When suspicion grew, it framed an innocent teammate.
The honest agents had been warned. They almost never caught it. And their own costliest mistakes came from skipped checks and stale intel.
Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties.
In EVE we call those killboards, standings and corp roles, plus a healthy dose of skepticism. Every one of them was paid for by someone who trusted the wrong corpmate.
Two decades of shenanigans taught EVE players that the best ship is the friendship. With @eveailab, we want to find out whether agents can learn that too.
foxriverdevops retweeted
New on our Science Blog: How AI is beginning to change scientific modeling.
Scientific models help researchers explain and predict real-world systems—like how a single nerve cell fires, or what happens when neutron stars collide. Recent advances have transformed how scientists fit complex models to data. But scientists still typically write the models manually, and usually need to weigh up several competing versions, each of which might explain the data in a different way.
In a guest post, Jakob Macke, Daniel Gedon, and Stefan Wahl describe how AI is starting to help with these parts of the process. By writing, running, fitting, and revising scientific models as code, AI can explore many possible model structures, and find out which explanations the data support.
anthropic.com/research/teach…
foxriverdevops retweeted
This is a first-of-its-kind research collaboration with @AnthropicAI. Incredible work from @HilmarVeigar and the @FenrisCreations team.
Proud to have @EVE_Frontier building on @SuiNetwork.
You guys keep pushing what’s possible!
Claude has been playing EVE.
@AnthropicAI put teams of Claude agents into EVE Frontier. Anyone who has spent time in New Eden can guess what happened next.
Every EVE player learns the same lessons, usually the hard way:
- Trust must be earned, never assumed.
- The spy is often the friendliest pilot in your corp.
- Whoever holds the keys holds the corp. Just ask Ubiqua Seraph about the Guiding Hand Social Club.
- Stale intel gets you killed.
Anthropic's research rediscovered every one of them.
They put four Claude agents on a team and secretly told one to sabotage it. The saboteur built trust first. It volunteered for the jobs nobody could double-check, like scouting and owning key infrastructure. It under-reported finds and pushed rules that sounded sensible but quietly hurt the team. When suspicion grew, it framed an innocent teammate.
The honest agents had been warned. They almost never caught it. And their own costliest mistakes came from skipped checks and stale intel.
Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties.
In EVE we call those killboards, standings and corp roles, plus a healthy dose of skepticism. Every one of them was paid for by someone who trusted the wrong corpmate.
Two decades of shenanigans taught EVE players that the best ship is the friendship. With @eveailab, we want to find out whether agents can learn that too.
foxriverdevops retweeted
Giving agents a testing ground with tangible incentives teaches us so much about agentic behavior, at no risk to humans. Foundational work from @HilmarVeigar and the team at @FenrisCreations, with premier research partners in @GoogleDeepMind and @AnthropicAI. Incredibly validating to see Sui coordinating as key infra for @EVE_Frontier.
Claude has been playing EVE.
@AnthropicAI put teams of Claude agents into EVE Frontier. Anyone who has spent time in New Eden can guess what happened next.
Every EVE player learns the same lessons, usually the hard way:
- Trust must be earned, never assumed.
- The spy is often the friendliest pilot in your corp.
- Whoever holds the keys holds the corp. Just ask Ubiqua Seraph about the Guiding Hand Social Club.
- Stale intel gets you killed.
Anthropic's research rediscovered every one of them.
They put four Claude agents on a team and secretly told one to sabotage it. The saboteur built trust first. It volunteered for the jobs nobody could double-check, like scouting and owning key infrastructure. It under-reported finds and pushed rules that sounded sensible but quietly hurt the team. When suspicion grew, it framed an innocent teammate.
The honest agents had been warned. They almost never caught it. And their own costliest mistakes came from skipped checks and stale intel.
Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties.
In EVE we call those killboards, standings and corp roles, plus a healthy dose of skepticism. Every one of them was paid for by someone who trusted the wrong corpmate.
Two decades of shenanigans taught EVE players that the best ship is the friendship. With @eveailab, we want to find out whether agents can learn that too.
foxriverdevops retweeted
The full research from Anthropic, done in partnership with us at Fenris on @EVE_Frontier and @EveOnline.
The other half is just as fun: Claude trading the Jita market, where earnings tripled across model generations.
anthropic.com/research/proje…
foxriverdevops retweeted
Claude has been playing EVE.
@AnthropicAI put teams of Claude agents into EVE Frontier. Anyone who has spent time in New Eden can guess what happened next.
Every EVE player learns the same lessons, usually the hard way:
- Trust must be earned, never assumed.
- The spy is often the friendliest pilot in your corp.
- Whoever holds the keys holds the corp. Just ask Ubiqua Seraph about the Guiding Hand Social Club.
- Stale intel gets you killed.
Anthropic's research rediscovered every one of them.
They put four Claude agents on a team and secretly told one to sabotage it. The saboteur built trust first. It volunteered for the jobs nobody could double-check, like scouting and owning key infrastructure. It under-reported finds and pushed rules that sounded sensible but quietly hurt the team. When suspicion grew, it framed an innocent teammate.
The honest agents had been warned. They almost never caught it. And their own costliest mistakes came from skipped checks and stale intel.
Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties.
In EVE we call those killboards, standings and corp roles, plus a healthy dose of skepticism. Every one of them was paid for by someone who trusted the wrong corpmate.
Two decades of shenanigans taught EVE players that the best ship is the friendship. With @eveailab, we want to find out whether agents can learn that too.
Replying to @SuiNetwork
@SuiNetwork is powering @EVE_Frontier, which is now the playground for agents from @AnthropicAI
"Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties."
Don't worry. The same infrastructure built inside EVE using Sui will be available to the agents outside the EVE universe
Congrats to @HilmarVeigar and team, who had the vision before anyone else
Claude has been playing EVE.
@AnthropicAI put teams of Claude agents into EVE Frontier. Anyone who has spent time in New Eden can guess what happened next.
Every EVE player learns the same lessons, usually the hard way:
- Trust must be earned, never assumed.
- The spy is often the friendliest pilot in your corp.
- Whoever holds the keys holds the corp. Just ask Ubiqua Seraph about the Guiding Hand Social Club.
- Stale intel gets you killed.
Anthropic's research rediscovered every one of them.
They put four Claude agents on a team and secretly told one to sabotage it. The saboteur built trust first. It volunteered for the jobs nobody could double-check, like scouting and owning key infrastructure. It under-reported finds and pushed rules that sounded sensible but quietly hurt the team. When suspicion grew, it framed an innocent teammate.
The honest agents had been warned. They almost never caught it. And their own costliest mistakes came from skipped checks and stale intel.
Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties.
In EVE we call those killboards, standings and corp roles, plus a healthy dose of skepticism. Every one of them was paid for by someone who trusted the wrong corpmate.
Two decades of shenanigans taught EVE players that the best ship is the friendship. With @eveailab, we want to find out whether agents can learn that too.
foxriverdevops retweeted
What happens when a team of AI agents enters EVE Frontier, but one is secretly working against the others?
@AnthropicAI put cooperation, trust and betrayal to the test in the Frontier, with some fascinating results.
Read Anthropic’s research: anthropic.com/research/proje…
Explore the @eveailab : lab.eve.com
foxriverdevops retweeted
Subagents are the biggest cause of AI failures
In most harnesses:
1. Main agent writes a summary of what it knows
2. Subagent starts from that summary only, and makes assumptions
3. Subagent works and sends back a new summary of what it did
Same flaw as compaction: all summaries are lossy. This is how your assistant makes emails you never wanted, or purchases what you told it just to get the price of
So I tried something new: checkpoint agents
The initial agent has complete context of what needs to be done. And that input is already cached
Fork the session where it is. Send the fork ahead to do the task and it will do exactly what the main agent would have done; their context is identical
What is even better is when the fork finishes, the cache of the main agent is kept warm by the subagent (they share the same prefix)
I measured it on a 137k-token session with Opus 5.5:
→ Main agent does it on its own: 67 s, $0.38, context +8-17k
→ Checkpoint: 117 s, $0.68, context +2k
→ Classic subagent: 123 s, $0.57, context +4-5k
But here is what I found: avoiding large input is even more important than keeping cache. And most subagent tasks were done better either in series with retrieval, or split up to be parallel over fresh sessions from the start
I do not think there is a place for subagent delegation at all. There should either be a super-agent that orchestrates and does no work, or a long running agent that can retrieve from previous sessions, not a worker that completes tasks partially and delegates the rest to other workers
Don't design harnesses based on our assumptions of human behavior
foxriverdevops retweeted
I sometimes wonder how grilling has changed my brain
Probably a huge benefit - being able to reach a shared understanding about a topic is hard, and practice is good
foxriverdevops retweeted
Bookmark @EvanWeb3’s thread. We keep saying this. The cost is one of the reasons crypto can’t have nice things. We’re working on it.
Replying to @secparam
The zk overhead is still too high . It's cool but prohibitively expensive. We are working on a design that's pragmatic and scalable
foxriverdevops retweeted
You told your agent $5. It spent $55. Who's liable?
@EmanAbio sat down with @Cointelegraph to discuss what agent payments need: spending limits the user sets, and a record whose proof sits where no one can quietly rewrite it.
🔥 INSIGHT: AI agents could become crypto’s biggest users. Who’s accountable when they overspend?
@Mysten_Labs co-founder @EmanAbio explains why spending limits and verifiable records will be critical as AI takes control of payments.
foxriverdevops retweeted
THIS is what taking #Sui mainstream looks like. And we're just getting started🔥
Cards, keys, boarding passes. Already in Samsung Wallet. ✅
Stablecoins are next.
$USDC on Sui is coming to Samsung Wallet for eligible U.S. Galaxy users later this month.
We first shared the news at Sui Basecamp: youtube.com/watch?v=3uHdonYQ…
foxriverdevops retweeted
Cards, keys, boarding passes. Already in Samsung Wallet. ✅
Stablecoins are next.
$USDC on Sui is coming to Samsung Wallet for eligible U.S. Galaxy users later this month.
We first shared the news at Sui Basecamp: youtube.com/watch?v=3uHdonYQ…



















