Joined June 2024
- Tweets 4,090
- Following 235
- Followers 378,758
- Likes 4,047
733 Photos and videos
Pinned Tweet
Jun 3
Today, we're going after one of AI's most important unsolved problems.
Introducing: Walrus Memory. 🦭
A portable memory layer that lets your AI agents carry context across every app you run them in.
No more starting from zero.
No more being locked into one platform.
Portable, verifiable, and fully under your control.
Take your agent's memory anywhere:
300
116
564
279,109
Walrus Memory plugs into the tools you're already building with.
→ Claude Code
→ Cursor
→ Codex
→ Gemini CLI
→ MCP (native support)
→ OpenClaw NemoClaw (first-party plugins)
→ Python, TypeScript, JavaScript SDKs
One curl command, no adapters needed.
curl -sL memory.walrus.xyz/skills/set…
Your agent has memory that survives the session.
🦭
29
39
140
4,217
There could be 100 AI agents for every person within a few years, and maybe a million each within ten.
@AlliumLabs talks about what that actually looks like, and why those agents are going to need verifiable data underneath them to function.
7
10
71
2,171
The model is rented, the memory is yours, and portability is the only thing that keeps it that way. 🦭
17h
every agent runs on the same models
what sets yours apart is its memory
THAT'S YOUR MOAT
so own it, and keep it portable @WalrusProtocol
6
15
81
3,805
Walrus 🦭/acc retweeted
Jun 14
when a government can force a model offline, you don't control your access to it
someone else decides whether you can access it or not
now imagine losing the context your agents build up every day, the memory they depend on
what's that worth to you?
if that memory sits inside the provider, it's locked in their system, and your access to it is decided by the same people who can pull the model
that's why we built memory.walrus.xyz
your agent's memory is encrypted and stored on @WalrusProtocol, spread across hundreds of nodes no single company owns, and you hold the keys
add it to the agent you're already running. it works with all the ai the tools you already use
56
66
369
14,262
Jun 13
Dear algorithm,
Please share this post with AI agent builders shipping in TypeScript or Python on LangGraph, CrewAI, and LlamaIndex.
They need their agents to remember what was learned in the last session, persist context across runtimes, and stop starting from zero on every new run.
🦭🙏
16
21
174
15,961
Mysten Labs (@Mysten_Labs) Co-Founder Kostas (@kostascrypto) argues that your AI memory could soon become more valuable than all of your historical online data combined:
"Agents are amnesiacs. You start a new session, they don't remember you, and they don't know your preferences."
"Everything your agent has learned from you, for your work, your preferences and your own secrets, is arguably more valuable than your data ever was in the past, including your social footprint, emails and search history."
"That's why we built memory around four principles: portability, verifiability, encryption and programmability."
"For the first time, people can have real cryptographic ownership of their digital identity rather than relying on passwords and fragmented accounts."
Jun 11
One truth just discussed in @etnshow
For years, people worried about what they posted on social media. Their photos, likes, comments, and social graphs became the product. The AI era changes everything:
For the first time, we're not just sharing our actions, we're sharing our thoughts, intentions, goals, fears, ideas, and reasoning. AI memory is becoming a digital extension of our minds, yet very few people think about who controls it, who can access it, how and when it's shared, or where it lives.
After spending many years at Meta as a lead cryptographer enhancing people's privacy, the more I think about it, the more I realize that AI memory may become even more sensitive than social media ever was...
That's why Walrus Memory is more important than even I originally believed.
7
25
115
11,350
Jun 12
A session ends. The context goes with it.
Walrus Memory keeps it. Two calls into your agent and the memory is still there next time you run it.
Learn more 👉 walrus.xyz/memory
8
19
116
4,896
Jun 12
"Data is the new gold."
@kostascrypto makes the case it has to be built without single points of control. 💭
12
23
116
5,693
Jun 11
Message to builders:
Build agents. Add memory. Repeat. 🔁🦭
9
17
112
4,453
Jun 11
Your AI memory is going to be more valuable than your Facebook, Google, or email footprint ever was. That's the bet @kostascrypto, Co-Founder and Chief Cryptographer at @Mysten_Labs, is making.
Catch him on @etnshow at 1:27:45 explaining how Walrus Memory makes that real. 🙏
Lovable $500M ARR, Fable 5, UK AI Minister Joins, OpenAI Price Cuts x.com/i/broadcasts/1pKdRRoeM…
18
18
107
4,853
Jun 11
Here's the bet: a few years from now, the hardest thing to copy about an agent won't be the model running it.
It'll be everything the agent has picked up from working with you over time.
The models keep getting cheaper. The memory built up around them keeps getting more valuable. 🦭
9
10
95
3,522
Jun 10
Walrus Memory mode is on.
Your agents finally get to remember. 🦭
26
33
204
5,446
Jun 10
"We're building something for the future."
@kostascrypto predicts the data we generate with AI today could eventually become our identity, or maybe even replicate us. 🦭
12
21
144
6,509
Jun 9
If you don't want to miss what Walrus is shipping, the newsletter's the place.
Subscribe 👇
walrus.xyz/newsletter/
18
16
149
6,061
Jun 9
Walrus Memory plugs into whatever you're already using. Works with Claude, ChatGPT, and Gemini, supports MCP natively, and ships with Python and TypeScript SDKs.
@EmanAbio walks through what it takes to actually start using it. 🦭
185
26
191
13,490
Jun 8
163 projects in two weeks is wild! 🔥
Big thanks to @tatum_io and to every team that shipped.
Winners are in the thread 👇
It turned into something much bigger than we expected 🤯
163 projects were built using during the Tatum x @WalrusProtocol Hackathon, 163!
A huge thank you to everyone who joined, built, experimented, and shipped.
And here are the winners 🏆
10
18
130
9,256
Jun 8
Builders, $70K is waiting! 🦭
Jun 5
Walrus 🤝 DeepBook 🤝 Sui Overflow 2026
Our two sponsors with specialized tracks.
9
8
101
7,588
Jun 8
Switching AI model providers right now feels like raising a new kid every time.
@AlliumLabs on the lock-in problem Walrus Memory was built to solve. 🦭
12
16
117
5,938
Jun 6
🚨 "Memory is one of the most critical bottlenecks in AI today.
Most agent memory lives locked inside platforms. Walrus Memory changes this. It puts builders in control and lets agents move and collaborate across different services. This is such an important foundation for the agentic future we all see coming."
— @kostascrypto, Co-Founder and Chief Cryptographer at @Mysten_Labs
17
31
188
7,751