Lead Data scientist | Blockchain amateur since 2019 | @SuiNetwork dreamer | SuiDrop 💧
Paris, France
Joined September 2015
- Tweets10K
- Following1.3K
- Followers2.1K
- Likes44.8K
Atalay 💧 retweeted
Last month, Meta, Google, Microsoft, Databricks, Couchbase, Cohere, and GitHub all rolled out new memory capabilities for AI agents.
We rounded up the biggest updates and what they signal for the AI industry. 👇
Atalay 💧 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…
Atalay 💧 retweeted
Agents will choose the fastest, cheapest, most robust ways to get things done for humans, and certainly when it comes to payments coordination, #Sui will be by far the most scalable and reasonable chain to do that on.
BULLISH🔥
Atalay 💧 retweeted
crypto or not, security is the one thing you have to obsess over rn
only the paranoid survive
Atalay 💧 retweeted
ICYMI: @github just made Copilot remember your preferences between sessions, even when it switches AI models mid-task.
That's great inside Copilot, BUT a lot of teams run agents in multiple places, and each of those agents still has to learn your preferences from scratch.
That’s the gap we’re building for: memory that goes with you across models and tools. ☝️
github.blog/changelog/2026-1…
Atalay 💧 retweeted
within the next 3 years, A2A payments will become orders of magnitude bigger than P2P payments
most of those payments will be in amounts too small, at speeds too fast and at a scale too big for any traditional rail we have today
#Sui is built for this
Atalay 💧 retweeted
agents are fundamentally non-deterministic by nature. you cannot prompt that away. so any agent you give open-ended spending power to will eventually do something you did not intend
non-determinism plus open authority is unbounded downside. you do not fix unbounded downside with better intentions. you fix it by making the downside bounded by design
that is the entire product thesis of VAA
Atalay 💧 retweeted
#Sui will become the place where agents coordinate and transact at scale
VAA is just the beginning
Atalay 💧 retweeted
One of the biggest stories from Sui Basecamp in Singapore: Verifiable Agent Arbiter (VAA), an evidence layer for AI agents being developed by @Mysten_Labs in collaboration with Google Cloud.
What was presented and why it matters 👇
Atalay 💧 retweeted
This is a great piece from @Forbes about Walrus Memory and VAA
I think people are really underestimating how much of AI is still up for grabs. There's a huge need for independent infrastructure that works across the big AI platforms
That's what we've been building at @WalrusProtocol and we've so much more to build from here
Featured in @Forbes: Walrus Memory.
“There was a point in time that we were stopped from accessing Fable and Mythos. And imagine at that time, what do I have to do?” — @kostascrypto
5,000+ agents and 4.5M+ retrievals later, here is why builders are moving to portable, user-keyed memory so switching LLMs doesn't mean starting from zero.
Read the full piece by @guneyyildiz: forbes.com/sites/guneyyildiz…
ALT If a company can swap one model for another in an afternoon, the real lock-in moves to everything its agents have learned about the business. Chalkias noted three reasons companies switch models: a better model arrives, a cheaper one does the job, or a regulator or employer restricts which systems can be used. He pointed to this summer’s interruption, between June 12 and July 1, when access to Anthropic’s Fable and Mythos models was suspended. “There was a point in time that we were stopped from accessing Fable and Mythos. And imagine at that time, what do I have to do?” Chalkias said. In most enterprise setups, switching models means leaving accumulated context behind. “You’re not able to go and get your data and move it to another LLM.”
Atalay 💧 retweeted
AI agents are getting smarter. The systems around them need to keep up.
Payments they can use. Data they can trust. Boundaries they can’t cross.
Five takeaways from Sui Basecamp on the infrastructure the agentic economy needs 👉 sui.io/blog/ai-lessons-from-…
Atalay 💧 retweeted
i believe people all over the world will use #Sui to move money every day. that's what we've been building for from the start. and that's what makes the partnership with samsung such a huge milestone
but this is only the beginning. we are collaborating on much much more
super pumped about what we can do from here
Atalay 💧 retweeted
Finding data across decentralized blobs and sites shouldn't require guesswork.
Join us on @discord for a builder AMA with @OmuraHQ co-founders Sourabh Chawla and Aditya Berry, breaking down real-time indexing and search infrastructure for Walrus.
October 21, 2026 at 12:00 p.m. UTC.
Bring your questions: discord.com/invite/walruspro…
Atalay 💧 retweeted
Using agents to explore search spaces and optimize thousands of models automatically.
@AnushElangovan of @AMD joined @ConorBronsdon to talk ROCm 10, agent swarms, and GPU performance.
Season 4 of @chain_ofthought sponsored by Walrus Memory: chainofthought.show/podcast/…
Atalay 💧 retweeted
A verified throughput record, a 99% cut to storage fees, and the registry behind 250 million AI training-data receipts moving onchain.
Here is your weekly recap of the core developments across the Sui network:
• @CertiK independently verified Sui's Basecamp live record of 40,682,328 peak TPS, with every tunnel opened and settled onchain.
• @CMEGroup's Q3 Crypto catch-up is out. The $SUI futures suite traded $253.5M in notional volume for the quarter, and SUI now sits in the CME CF Emerging Crypto Index alongside the nine other largest assets outside bitcoin and ether.
• @datafdn's DATA Network and Trace, an onchain registry with 250M+ verifiable receipts, are joining the Sui stack. Walrus becomes the default storage layer, so AI labs get one place to source real-world training data they are cleared to use, with the provenance regulators now ask for.
• A protocol upgrade cut non-refundable storage fees by 99%. @AftermathFi estimates 60 to 70% less total gas on quoting transactions, which makes placing, cancelling, and updating orders cheaper for market makers on a fully onchain orderbook.
• Aftermath listed DATA perps. Long or short with up to 3x leverage.
• @SlushWallet integrated @HallidayHQ. Bring tokens from 25 chains, or fund with a card, bank transfer, or Apple Pay, and receive $SUI or $USDC in a few taps.
• @realtbook released episode 2 of fin technica. Co-founder Nick and @Mysten_Labs Head of Stablecoins @louisabai on the economics of stablecoins and the race to control their distribution.
• @Grayscale Head of Research @LowBeta on the Basecamp stage: "SUI is building the infrastructure for crypto's next chapter."
• @PredictBay shared its full Basecamp talk on where prediction markets go next. Built on Sui with @DeepBookonSui and @WalrusProtocol.
• @ONEChampionship ran ONE SAMURAI 3 with Sui as its official blockchain partner. Yuki Tanaka finished Aiji Kobayashi in 67 seconds, September's fastest finish.
Atalay 💧 retweeted
Welcome to SUI 101.
Spend a few minutes learning about @SuiNetwork (SUI) in the app and get $5 in SUI.
Rewards while supplies last. Not available in NY state. Other terms apply.
Atalay 💧 retweeted
Building an AI agent? Give it memory that survives the session.
If you're tired of re-explaining context or losing memory between runs, here's how to get started with Walrus Memory:
Core SDK: mysten-incubation/memwal stores encrypted memory on Walrus, with independently verifiable data integrity.
github.com/MystenLabs/MemWal…
OpenClaw plugin: mysten-incubation/oc-memwal adds automatic memory capture and recall to OpenClaw agents, keeping memory available across sessions.
github.com/MystenLabs/MemWal…
AI-friendly docs: SKILL.md provides integration instructions, while llms.txt gives AI tools a structured documentation reference.
github.com/MystenLabs/MemWal…
Persistent memory across sessions, tools, and workflows. Under your control.
Quickstart + source code: github.com/MystenLabs/MemWal