nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul) header
nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul) profile picture

Yaping where it’s hot, staying onchain where it counts. #GoodThingsTakeTime

indonesia
Joined February 2017
Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)Photo by nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) (@direumatteul)
something i find interesting about humanoid robots is that they’re being built to operate in a world that was never designed for them doors, stairs, elevators, tools, and even simple everyday objects are all made around human movement and habits so having two legs and two arms isn't enough. a robot still needs to understand how to interact with all these things and adapt when situations get tricky imo, the real breakthrough won't just be robots that look more human. it'll be robots that can actually function in our environment without needing everything to be redesigned for them that's one of the challenges worth watching as physical ai develops, and why i'm keeping an eye on @continuumlabs_
12
14
163
one interesting detail about @deloraprotocol is how it handles execution at the smart contract level its Diamond Contract uses a modular architecture with different Facets for different execution tasks depending on the selected route, the contract can use the appropriate Bridge or DEX Facet to process the transaction i think this is an important part of building flexible multichain infrastructure. different routes can require different execution methods, so having dedicated components makes the architecture more organized it's not just about finding a route with good liquidity. the way that route gets executed matters too
5
1
10
121
one thing i've been paying attention to with @get_optimum is not just the technology itself, but how adoption is starting to grow recently, Optimum shared that more than 3 million ETH, nearly 8% of all staked ETH, is now running mump2p that's actually pretty interesting when you think about it. building a faster networking protocol is one thing, but getting real validator operators to integrate and use it is another challenge we're also seeing more teams like Puffer and FP Validated testing the protocol, which makes the ecosystem worth watching for me, this is an important part of Optimum's progress. better block propagation brings technical improvements, but growing adoption is what can help those improvements reach more of the Ethereum network still early, but seeing more operators get involved is a good sign. curious to see how far this can grow from here
8
9
146
been looking into another interesting part of Fluton's privacy infrastructure this time it's about EIP-712 deposit screening @FlutonIO has updated its Privacy Pool mechanism to use signed deposit screenings, adding another important layer to how users are checked before entering the pool what caught my attention is the follow-up review by Kann Audits. they looked at signature verification, expiry handling, replay protection, and signer rotation these details might not look exciting at first, but they're important when building privacy infrastructure that people can actually rely on privacy is not just about hiding transactions. the systems behind deposits and verification need careful attention too
13
1
15
179
moving assets across different chains is one thing, but keeping track of every transfer is just as important this is another feature i find interesting about @deloraprotocol through its API, developers can quote, swap, bridge, configure their own fees, and track every transfer across EVM networks and Solana having transfer tracking built into the infrastructure makes it easier for applications to follow what's happening behind each transaction developers don't just need a way to move funds. they also need visibility into the process that's an important part of building a better multichain experience
34
42
329
privacy and compliance don't always have to be on opposite sides that's what makes compliance aware privacy pools interesting to me users want to keep their transfers and financial activity private, but protocols still need ways to screen deposits and manage risks @FlutonIO is exploring this through its Privacy Pool architecture, including Predicate deposit screening the recent Kann Audits review also looked at how this component connects with the wider pool system i think this is an important direction for defi. privacy should protect user activity without ignoring the security and compliance requirements of the ecosystem finding that balance is what makes onchain privacy worth building
11
21
198
something interesting is happening in web3. traditional financial markets are slowly moving onchain, and tokenized equities are becoming a bigger part of the conversation but there is one thing i think people should pay more attention to: latency in trading, getting market information earlier can make a real difference. traders need to react quickly, especially when prices move and opportunities do not stay around for long this is where @get_optimum caught my attention their Universal Data Acceleration Network is designed to make data move faster across decentralized networks. and now they are looking at how this technology could support onchain equity trading too what makes this interesting is that onchain markets have different infrastructure challenges from traditional trading. simply bringing stocks onto a blockchain is not enough. the systems delivering market data also need to keep up imo, the future of onchain finance will depend on more than liquidity and execution. reliable, low latency data delivery could become just as important
18
22
225
been checking out the latest demos from @4Dlabs_Official, and there's something interesting about what they're building for robots to learn from real human behavior, capturing a scene is only the beginning. the raw data also needs to be cleaned, o rganized, and properly annotated before it can become useful training data. for robots to learn from real human behavior, capturing a scene is only the beginning. the raw data also needs to be cleaned, organized, and properly annotated before it can become useful training data. that's why i find the embodied data pipeline particularly interesting 4D Labs is working on connecting first-person capture hardware with data cleaning and annotation workflows to turn real human actions into structured data for robotics training i think this is an important part of physical ai that doesn't get enough attention better robots won't come from better models alone. they also need better real world data to learn how humans interact with the world still early, but i'm curious to see how far 4D Labs can take this
13
21
231
speed really matters when moving assets between different chains one thing i find interesting about @deloraprotocol is its latest progress on cross chain settlement speed according to their latest update, settlement time has gone from 87 seconds to just 2 seconds in a month, with cross-chain routes now averaging 40x faster that's a pretty meaningful improvement for anyone building apps that depend on moving liquidity across networks users don't want to spend extra time waiting for their transactions to finish. they just want to move their assets and keep using the app making cross-chain transfers feel closer to a regular swap is definitely a direction worth watching good to see @deloraprotocol putting real focus on execution speed, not just adding more chain integrations
10
17
191
one thing that caught my attention about ethereum's upcoming Glamsterdam upgrade is how ePBS could change the whole data delivery game we usually talk about block builders, MEV, and better bids, but there's another challenge happening after a bid gets selected: getting the actual execution payload to the network on time under ePBS, this becomes even more important. the winning builder needs to deliver its payload to the 512 validators in the Payload Timeliness Committee before the 9 second deadline think about it. a builder could win a valuable bid, but if the payload doesn't reach the network in time, that opportunity could be wasted this is where @get_optimum gets interesting mump2p is built to make P2P data propagation faster and more efficient, and Optimum is preparing to bring that capability to the builder market as Glamsterdam approaches imo, blockchain performance isn't just about building better blocks anymore. getting the data to the right participants at the right time matters just as much with ePBS, milliseconds could become even more valuable
20
21
219
building a privacy product is one thing, but making sure people can actually use it smoothly is another i've been seeing more activity around @FlutonIO is private testnet, including community guides and updates to its role progression that's a good sign of how important community participation can be during the early stages of building testing, exploring the features, and sharing experiences all help make the journey better with the trusted setup already completed by 222 contributors, i'm curious to see how the next phase develops still early, but definitely worth keeping an eye on Fluton
21
26
254
nat die (! π•œπ•’π•žπ•’π•π•«) | (❖,❖) retweeted
something i find interesting about FHE is that speed isn't the only thing that matters in blockchain, getting the same result across different nodes is just as important. you don't want each node coming up with a different output when processing the same confidential operation this is why one detail in @primus_labs AutoHoG caught my attention everything runs using integer arithmetic, allowing nodes to produce byte identical results imo, deterministic computation is a pretty important piece of confidential onchain finance. privacy means little if the underlying system can't produce consistent and reliable results AutoHoG is tackling both performance and consistency, which makes the direction worth watching i think this is the kind of technical progress that can make FHE more practical beyond just research explore primus xp: s.kaito.ai/Pp4CQVu
34
2
33
376
a trusted setup is one thing, but being able to look back at how the process happened matters too what i find interesting about @FlutonIO is that the ceremony doesn't just end with 222 contributors and a completion announcement there's also a public transcript showing contributions in order, along with their hashes and zkeys i like this approach because transparency matters when building privacy infrastructure. people should have a way to inspect the process, not just hear that everything went well trusted setup complete. now it's time to see what comes next with beta mainnet
222 contributors, trusted setup complete. now the next chapter for @FlutonIO begins beta mainnet is what i'm looking forward to next the interesting part now is seeing how all the privacy infrastructure comes together in practice. from encrypted transactions and confidential execution to private transfers and settlement, there's a lot to watch as Fluton moves into its next phase the recent Kann Audits review also covered important components like the Privacy Pool, Merkle tree, Groth16 verifiers, Circom circuits, and FHE wrappers pretty exciting progress so far. curious to see what the beta brings and how the community gets involved next
19
24
200
something i've been thinking about lately is how latency can actually create economic value in web3 we usually talk about network speed as a technical metric, but for validators and traders, a few milliseconds can make a real difference. getting information earlier can mean more time to evaluate bids, react to market activity, and make better decisions this is where the idea of a Latency Marketplace gets interesting @get_optimum is exploring how to turn the speed advantages created by RLNC into something with economic value. the idea is that different participants can benefit from different levels of data delivery speed, depending on what they need and how much they value that extra time what i like about this concept is that it connects networking technology with market economics imo, speed shouldn't only be an advantage for participants with expensive hardware or the best geographic location. improving the network itself could make low latency more accessible across decentralized systems there's a lot more to this than simply making blocks arrive faster. and that's what makes Optimum worth following
14
19
228
something i find interesting about FHE is that speed isn't the only thing that matters in blockchain, getting the same result across different nodes is just as important. you don't want each node coming up with a different output when processing the same confidential operation this is why one detail in @primus_labs AutoHoG caught my attention everything runs using integer arithmetic, allowing nodes to produce byte identical results imo, deterministic computation is a pretty important piece of confidential onchain finance. privacy means little if the underlying system can't produce consistent and reliable results AutoHoG is tackling both performance and consistency, which makes the direction worth watching i think this is the kind of technical progress that can make FHE more practical beyond just research explore primus xp: s.kaito.ai/Pp4CQVu
34
2
33
376
one thing that makes @deloraprotocol interesting is how it's expanding access to different blockchain ecosystems with Delora Trade, users can now bring tokens from 30+ chains directly into @Enter_Elysium what i like about this is that users don't have to think too much about where their assets are coming from. they can access Elysium from different networks through one interface, while Delora handles the routing to find a suitable path this kind of connectivity makes an ecosystem more accessible and gives users more flexibility when moving between chains more connections, more opportunities for apps and users
14
14
223
never really thought about how much information goes into something as simple as holding a paper cup grip it too lightly and it slips. squeeze it too hard and it gets crushed. finding the right amount of pressure is something humans do naturally, but teaching a robot to understand that is a whole different challenge this is why i find what @4Dlabs_Official is working on pretty interesting capturing vision, motion, and touch during real-world tasks could give robots more useful information about how humans interact with objects and their surroundings imo, physical ai needs to learn more than just what things look like. it also needs to understand how much pressure to apply, how objects respond, and how to adjust its actions in different situations turning these real interactions into training data feels like an important step toward making robots more useful in everyday life definitely an interesting direction to follow
Grip a paper cup too lightly and it slips. Squeeze too hard and it buckles. At #4DLabs, we're building tools to capture vision, motion and touch during real tasks, then turn those signals into training data for robots.
9
1
15
233
222 contributors, trusted setup complete. now the next chapter for @FlutonIO begins beta mainnet is what i'm looking forward to next the interesting part now is seeing how all the privacy infrastructure comes together in practice. from encrypted transactions and confidential execution to private transfers and settlement, there's a lot to watch as Fluton moves into its next phase the recent Kann Audits review also covered important components like the Privacy Pool, Merkle tree, Groth16 verifiers, Circom circuits, and FHE wrappers pretty exciting progress so far. curious to see what the beta brings and how the community gets involved next
222 contributors. one completed trusted setup ceremony. now the next chapter begins pretty cool to see the @FlutonIO community come together and take part in something this important a trusted setup is not just another milestone. it's part of building the cryptographic foundation behind privacy infrastructure, and having community contributions makes the process even more meaningful now the focus shifts to beta mainnet curious to see how Fluton takes this next step. the journey from building the privacy foundation to bringing it closer to real onchain use is getting interesting big milestone for the team and everyone who contributed
13
1
18
392
one interesting challenge in blockchain networking is how much data a single source node has to send to other participants as the number of validators grows, the source node can become a bandwidth bottleneck. sending data to more peers means more pressure on its outgoing capacity, which can affect how efficiently blocks are delivered this is where Source Bandwidth Optimization gets interesting @get_optimum is exploring this through Peer-Turbo, using RLNC to let receiving nodes exchange coded data with each other instead of relying entirely on the original source what i like about this approach is that it looks at the network as a whole, not just the speed of one node their research also explores how this can work with blockchain networks like Solana and Monad, not just Ethereum imo, making decentralized networks more efficient isn't always about adding more bandwidth. sometimes, improving how participants share data can make a real difference too
21
26
334
been thinking about the biggest challenge with privacy onchain, and imo it is not just about keeping data hidden it is also about making encrypted computation fast enough to be useful traditional FHE can become pretty expensive when a simple operation gets broken down into thousands of gates, with bootstrapping adding another layer of computational overhead this is where AutoHoG from @primus_labs caught my attention instead of treating every gate separately, AutoHoG automatically designs custom homomorphic gates and combines mergeable subcircuits to reduce bootstrapping work according to Primus, a 64-bit confidential transfer can now take around 90ms, reaching 210+ transfers per second across 8 RTX 5090 GPUs what i find interesting is the direction behind this privacy infrastructure should not force users to choose between keeping their financial data confidential and having a system that actually performs well FHE optimization might not be the flashiest part of web3, but it could be one of the things that makes confidential onchain finance more practical worth keeping an eye on what Primus builds next explore primus xp: s.kaito.ai/Pp4CQVu
something i think web3 still needs to figure out is how to trust information that comes from outside the blockchain blockchains can verify onchain activity, but a lot of important information still lives on websites, APIs, and traditional platforms think about prediction markets or private credit. the data behind these systems needs to be reliable, but users shouldn't have to expose all their personal information just to prove something is true that's why zkTLS caught my attention what @primus_labs is exploring around verifiable offchain data is pretty interesting. the idea of proving that information is authentic while preserving user privacy could open up more possibilities for onchain finance imo, verification should make information more trustworthy, not force people to give up their privacy this is the kind of infrastructure web3 needs to mature explore primus xp: s.kaito.ai/Pp4CQVu
19
1
26
257
privacy only becomes useful when it can keep up with real onchain activity. 90ms for a confidential transfer is a pretty interesting step.
2
22