Comms @mysten_labs, formerly @atptour. Highlighting elite communicators. All views my own.
Joined June 2013
- Tweets609
- Following2.8K
- Followers652
- Likes3.9K
Mark Epps retweeted
1/ Sharing the first update in a new quarterly series from @Mysten_Labs – take a look mystenlabs.com/blog/letters-…. In it, I lay out my thoughts on:
What problems need to be solved in order for agents to be trusted by humans and to trust each other.
Why I’m convinced Sui and Walrus will be fundamental in solving those problems, and why Mysten Labs is the team to tackle them.
2/ We’re also setting up a new, dedicated AI business unit within Mysten Labs, which I will personally lead, in addition to my responsibilities as CTO. At Sui Basecamp, I gave a glimpse of what we’re working on: youtube.com/watch?v=rCWnarkg…
3/ To ensure that the team has full time and bandwidth to dedicate to our agentic strategy as well as supporting the Sui and Walrus ecosystem, I have asked @KevinBoon_ML to step into the CEO role. Kevin has been instrumental in leading Mysten Labs alongside the co-founders from Day 1, first as our Chief Legal Officer and then as our President. I planned for this day when I made him President, and have the utmost faith in him as my successor and our new CEO.
Now, back to work.
Mark Epps retweeted
In collaboration with @googlecloud we’re announcing Verifiable Agent Arbiter (VAA), an evidence layer enabling enterprises to prove their AI agents acted as authorised, using @SuiNetwork and @WalrusProtocol to anchor that proof 👇
Mark Epps retweeted
Reportedly working on affordable Quantum 2FA Sui cards.
What happened to Coldcard will NEVER happen to my people, but I need to step up with personal time (dear wife, please excuse me)
Last year, with help from some friends, I leased a dedicated factory in a secret location to build quantum-safe wallets at scale. Target: < $10 per quantum card key and < 1-2secs per NFC quantum signature.
Crypto users shouldn’t be the victims. I won’t allow it.
And yeah, I might even sponsor them for the very poor. Everyone deserves top-tier protection hygiene.
We shouldn’t be reacting to the quantum or bad randomness threat. We should lead it.
And hopefully, one day, even the banks will be using our tech, as you all know me… I’m devoted 🫡 to crypto!
Mark Epps retweeted
Auto-evolved addresses.
Same wallet mnemonic with automatic signature lift, afaik our yesterday’s discovery in the Sui cryptographers’ meditation 🧘 gathering, revealed a new algorithm. More soon
Mark Epps retweeted
LFG!!
Mark Epps retweeted
"It's been one of the biggest upgrades for my productivity."
Huge thanks to @nateherk for breaking down how Walrus Memory lets you jump between Claude and ChatGPT without ever having to re-explain yourself.
Check out the breakdown on YouTube:
youtube.com/shorts/NClayXM8p…
Mark Epps retweeted
We used to share our lives on social media. Now we share our secrets with LLMs.
@Mysten_Labs co-founder @kostascrypto, speaks on why your chat history with AI is more revealing than anything you ever posted.
Mark Epps retweeted
Data is the expensive, unsolved part of building #AI agents.
@kostascrypto shares why Walrus Memory is one of the strongest attempts yet at fixing it.
What’s breaking in your agent’s memory setup?
Mark Epps retweeted
We've raised $1.6M to build Europe's first daily live technology show and we're live right now.
State-of-the-art newsroom near Kings Cross.
First guests being announced shortly.
Coming to you 5 days a week.
What a time to be live.
12-3 GMT.
Mark Epps retweeted
etn has raised $ 1.6M. From today, etn will be live 5-days a week out of our new 1500sqft studio in central London.
In 10 months Ronan and I have scaled etn from just us 2, to 8 people. We’re building a team of people who feel as strongly as we do about building the tech network the continent needs.
Our chief of staff, Alex said this to me this morning before we went live: Today will be the first time in history a company is live streaming the daily growth of a company.
1. Companies globally are being built faster than ever in history. Europe recently saw two of its largest ever seed rounds with AMI Labs and Ineffable Intelligence. @ElevenLabs just crossed £600M ARR in under 2.5 years, @WeAreLegora achieved one of the fastest growth trajectories in enterprise software history.
2. Technology is converging with culture. Technology entrepreneurs are becoming celebrities, celebrities are building AI companies. E.g. David Beckhams IM8 just raised $ 1B. We’ve experienced this in real time with the spread of guests on etn: F1 champions, world renowned authors, Government Ministers, public CEO’s, frontier AI Lab founders.
3. Everyone has a point to make in the global AI conversation. For AI to diffuse into everyday life, in schools, hospitals, manufacturing etc, the conversation needs to two things: Happening all the time and positive. etn is the channel from Europe that distributes that conversation. Read our manifesto on our website.
4. X is the global newspaper where everyone has the ability to distribute and own their narrative. Recently explained by @jeremygiffon on @patrick_oshag's Invest Like The Best I felt this put it perfectly. The distribution of news is taking on new forms - people want to hear directly from other people, speed is critical because of how fast industries are moving.
We’re live now and accepting guests.
Details on our website - etnshow[dot]co
What a time to be live.
L & R
Mark Epps retweeted
SK Hynix is about to raise $28 BILLION through its NASDAQ IPO this Friday
This is arriving at a moment where investors are finally treating memory as core AI infrastructure, as opposed to a secondary hardware trade
That same logic should eventually extend into crypto
As agents become more useful, the demand shifts toward "verifiable" storage, persistent context & programmable memory layers...
These are properties crypto ACTUALLY does well
If you want to learn more, then our @KhalaResearch report on Walrus sits directly in that gap
It’s now obvious that memory is one of the biggest bottlenecks in AI
But the next question is whether that memory can be trusted
“Context is the new bottleneck” - Jensen Huang at CES 2026
He’s right, but only half right
Every AI agent deployed today still has the memory of a goldfish because context windows are stateless by design, starting each session from zero
That is unfortunately how LLMs work
But storage capacity is only half the battle, the other half is trust
Current memory providers can write memories to docs, databases or internal logs, then ask enterprises to trust that record later
The problem is obvious: one altered memory corrupts every decision that relied on it
As agents move into regulated workflows, this becomes a real issue
The EU AI Act requires high-risk AI systems to maintain automatic event logging and traceability, meaning enterprises will need proof of what data an agent accessed, what it remembered and whether that record was tampered with
Our latest report covers AI memory and how Walrus tackles this problem with portable, verifiable and programmable memory for agents
The full report is in the next post below
Mark Epps retweeted
I’ve been digging deeper into the memory trade in crypto
Micron, SK hynix, Samsung etc have already had Goliath runs off the AI memory thesis, but the crypto version of the trade still feels under-researched and barely repriced
Our Khala post below explains why this matters; AI agents don’t just need bigger context windows, they need memory that can be trusted, verified and carried across platforms
That is why Walrus Memory is interesting; we sat down with the Mysten Labs (Sui) + Walrus teams to understand how onchain memory can extend the AI memory stack with verification, portability, programmable access and provenance
While everyone is chasing equities, I think the asymmetric research is still hiding in beaten down crypto
Time to pay attention, particularly given the valuation gap:
It’s now obvious that memory is one of the biggest bottlenecks in AI
But the next question is whether that memory can be trusted
“Context is the new bottleneck” - Jensen Huang at CES 2026
He’s right, but only half right
Every AI agent deployed today still has the memory of a goldfish because context windows are stateless by design, starting each session from zero
That is unfortunately how LLMs work
But storage capacity is only half the battle, the other half is trust
Current memory providers can write memories to docs, databases or internal logs, then ask enterprises to trust that record later
The problem is obvious: one altered memory corrupts every decision that relied on it
As agents move into regulated workflows, this becomes a real issue
The EU AI Act requires high-risk AI systems to maintain automatic event logging and traceability, meaning enterprises will need proof of what data an agent accessed, what it remembered and whether that record was tampered with
Our latest report covers AI memory and how Walrus tackles this problem with portable, verifiable and programmable memory for agents
The full report is in the next post below
Mark Epps retweeted
I invited @Mysten_Labs co-founder Kostas onto the Supercycle to discuss the state of AI
Kostas has spent his career deep in cryptographic research, led cryptography at Meta and has spent the last five years at Mysten Labs delivering research and shipping applications for developers
This guy is one of the most credible in the world on cryptography
We spoke about privacy tech, the rise of agents in enterprises and the major bottlenecks for crypto adoption
Timestamps:
00:00 Introduction
01:33 The Mysten Labs Thesis
04:10 Silicon Valley’s thesis on decentralisation
08:42 Agentic Economics
13:30 Delivering global blockchain adoption via enterprises
18:17 Crypto’s obsession with privacy
24:39 Solving Agent Memory
37:59 Decentralisation vs AI power concentration
#sui #walrusmemory #crypto
Mark Epps retweeted
It’s now obvious that memory is one of the biggest bottlenecks in AI
But the next question is whether that memory can be trusted
“Context is the new bottleneck” - Jensen Huang at CES 2026
He’s right, but only half right
Every AI agent deployed today still has the memory of a goldfish because context windows are stateless by design, starting each session from zero
That is unfortunately how LLMs work
But storage capacity is only half the battle, the other half is trust
Current memory providers can write memories to docs, databases or internal logs, then ask enterprises to trust that record later
The problem is obvious: one altered memory corrupts every decision that relied on it
As agents move into regulated workflows, this becomes a real issue
The EU AI Act requires high-risk AI systems to maintain automatic event logging and traceability, meaning enterprises will need proof of what data an agent accessed, what it remembered and whether that record was tampered with
Our latest report covers AI memory and how Walrus tackles this problem with portable, verifiable and programmable memory for agents
The full report is in the next post below
Mark Epps retweeted
Big congrats to our partner @alliumlabs on the $40M raise 🎉
Turning onchain data into something institutions can use is hard, and they’re one of the few teams pulling it off.
Pumped to keep building together. 🦭
Mark Epps retweeted
Your AI agent's memory is probably stuck inside one app.
Walrus Memory (@WalrusProtocol) fixes that with an MCP that creates memories as you chat, then can recall them anywhere - Claude Code, Pi, etc.
Portable and actually owned by you!
Mark Epps retweeted
.@Cole_Medin used a single prompt to set up Walrus Memory across Claude Code, Codex and Pi, and it stays encrypted and owned by him the whole way.
Watch the demo 🦭👇
Your AI agent's memory is probably stuck inside one app.
Walrus Memory (@WalrusProtocol) fixes that with an MCP that creates memories as you chat, then can recall them anywhere - Claude Code, Pi, etc.
Portable and actually owned by you!