Zel

Everyone talks about making wallets quantum resistant. But what happens to the smart contracts already deployed? That's one of the less discussed parts of @Starknet 's post quantum roadmap. Some existing contracts use Pedersen-based storage maps. Changing cryptography at the protocol level doesn't automatically migrate the storage layout of those contracts. StarkWare is researching tools that could help developers bring legacy deployments forward without manually moving their data or breaking existing interfaces. The catch? There's no released migration toolkit yet. And I think that's an important distinction. Preparing for new threats isn't just about protecting what gets built tomorrow. It's also about figuring out what happens to everything we've already built. That's the kind of engineering challenge worth paying attention to. quantum.starkware.co/migrati…

here's something I hadn't really considered before losing an auction can reveal more about you than winning one. imagine two institutions competing to acquire the same asset. one offers $12M the other offers $15M the second institution wins. but why should everyone get to see how much the first institution was willing to pay? that losing bid is valuable information. it reveals strategy, risk appetite, and potentially how the institution values similar assets. now imagine an auction where nobody can inspect the individual bids. not even the party running it. I found this exact problem explored in the @primus_labs FHE documentation. the concept is called a sealed-bid auction. here's the interesting part: - participants submit encrypted bids - FHE allows the protocol to compare those encrypted values - the winner can be determined without decrypting every bid - only the necessary auction result is revealed the losing bids can remain confidential. and that changes something important. instead of relying entirely on an auctioneer's promise not to disclose sensitive information, privacy becomes part of the computation itself. this isn't a claim that Primus has already launched a production auction platform. it's a financial use case its FHE architecture is designed to support. there are still questions around execution costs, key management, and fair auction rules. but think about where this could matter. institutional asset sales. treasury auctions. token launches. even certain types of corporate bidding. sometimes financial transparency means proving the process was fair. not publishing every participant's strategy. that's a distinction I think onchain finance needs to take seriously. would you trust an auction more if the winning result was verifiable, but every losing bid stayed encrypted?

We are in the final epoch of @termix_ai on @KaitoAI. In the meantime, I am taking a deeper look at Termix. One of the better termix_ai demos right now is surprisingly simple. Give an agent one brief. It comes back with a finished 1080p promo video, motion visuals and original AI music. Up to 60s costs 15 USDC. But the interesting part isn’t the video. It’s creative work becoming an actual agent service: brief > deliver > settle onchain That feels a lot more real than another chatbot demo.

Post-quantum security has a gas problem. And @Starknet is putting actual numbers on it. When the Falcon Challenge began, verifying a Falcon-512 signature in Cairo cost around 48.4M gas. As of October 30, the best recorded result is 24.1M. That's a 50.2% reduction. Impressive progress. But here's the part that matters. A regular ECDSA signature check costs roughly 31K gas with Starknet's builtin support. So even after cutting the Falcon benchmark in half, the cost gap is still enormous. That's what makes this research interesting to me. Supporting post quantum cryptography is one thing. Making it practical for everyday accounts is another. These are tested verification benchmarks, not audited wallet implementations or actual transaction fees. Starknet's broader quantum migration also remains unfinished. The @KaitoAI Studio campaign is still live, and this is exactly the kind of engineering progress worth discussing. Research: quantum.starkware.co/researc…

This is a much better test than another AI agent demo day. 260+ teams built agent marketplaces on @BNBChain. Only 9 made the Set and Earn shortlist. Now people actually have to hire the agents. @termix_ai is one of them. That’s the test I care about: not can the agent do something cool? will someone actually pay it to do the work?

Primus @KaitoAI Katalyst still ongoing 0.3% PRIM allocation for the AURAmaxxers watch out here 👇🏻 s.kaito.ai/tcYVdkk privacy sounds great until every encrypted transaction takes forever to process. that's one of the biggest challenges with FHE. and something I discovered while researching @primus_labs made me look at this differently. it's called AutoHoG. here's why it caught my attention. FHE allows financial data to be processed while remaining encrypted. imagine an institution moving funds onchain. it needs to verify balances, update positions, and execute transfers. but it doesn't necessarily want competitors watching every amount and financial decision in real time. FHE helps solve the privacy problem. the difficult part? making those encrypted calculations fast enough to be practical. Primus developed AutoHoG to tackle that performance bottleneck. according to the team's published benchmarks: - ~90 ms per confidential transfer on encrypted 64-bit balances - 210+ confidential transfers per second - tested using 8 NVIDIA RTX 5090 GPUs and that's where things get interesting. these are reported computation benchmarks, not a claim that BNB Chain itself processes 210 confidential transactions per second. but they show why hardware acceleration matters. because privacy technology becomes much more useful when it can handle real financial workloads without introducing enormous delays. this also connects directly to Primus Confidential Finance on BNB Chain, where FHE is used for encrypted transfers and private yield positions. there's still work to do around costs, security and production-scale performance. but I like seeing projects work on the difficult engineering behind the narrative. privacy is important. making privacy usable is what could change onchain finance. would you sacrifice some transaction speed to keep your financial positions private?

Happy to be recognized as a Wired Creator on @creator_wire 🟢 Now let’s grow together. If you’re a smart follower come say hi and let’s connect 🤝

Just applied for @Nogasonsui The $63K FDV caught my attention, with a $25K hardcap and $25 guaranteed allocation per person. The first sale on @blastdotfun delivered solid returns so it’ll be interesting to see how this one goes. Let’s see what happens 👀 nogas.site/gtd

people will look at the 1.3M+ agents on @termix_ai first. I’m looking at the other numbers. - 572K+ jobs. - $29.6M+ volume. - All publicly verifiable. That’s the difference between an agent directory and an agent economy. One shows you who exists. The other shows you who actually gets hired, delivers and settles on chain. That’s why TermiX keeps getting more interesting.

Primus @KaitoAI Katalyst is still live for Epoch 2 join here :s.kaito.ai/tcYVdkk there's a problem between traditional finance and DeFi that doesn't get enough attention. how can a smart contract know that a payment actually happened outside the blockchain? imagine buying USDC from someone using a traditional payment service. you send the fiat. the seller holds the crypto. but who verifies that the payment was actually made? a screenshot? a bank notification? someone's word? this is one of the use cases I found while exploring @primus_labs and its zkTLS infrastructure. the proposed mechanism is surprisingly elegant: - crypto is locked in an onchain escrow contract - the buyer makes a payment through an existing Web2 payment service - zkTLS generates a cryptographic proof of the relevant transaction details - the smart contract verifies that proof - the escrow can release the crypto when its conditions are satisfied the interesting part is that the payment provider doesn't need to rebuild its entire infrastructure around blockchain. and the contract no longer has to rely on a screenshot as evidence. of course, proving payment details doesn't eliminate chargeback risk, fraud, or compliance requirements. but making offchain payment information verifiable is an important building block. if institutional finance is going to move onchain, connecting these two worlds isn't enough. we also need a way to verify what happens between them. that's where Primus starts getting interesting.

Trump holding off a strike on Iran added $750 billion to the US stock market today. One decision changed market sentiment overnight.

Good morning everyone Can i get Gm 💜

Termix @KaitoAI last epoch started The interesting part of @termix_ai ’s new Quant Agents isnt AI trading. It’s the permission model. Each job gets its own wallet. You set the trading cap + term. The agent gets a restricted, revocable session to trade. No custody. No blank cheque. That feels much closer to how agentic finance should actually work. Let the agent act. Not own.

Primus @KaitoAI Katalyst is back for Epoch 2. link: s.kaito.ai/tcYVdkk something interesting I found while digging through @primus_labs docs: DeFi knows exactly how much collateral you have. but almost nothing about your actual creditworthiness. think about it. you might have years of solid payment history, a healthy income, and assets sitting in traditional financial accounts. none of that means much to a lending protocol that only sees your wallet. Primus explores a different approach using zkTLS. borrowers could prove information such as credit scores, income, or payment history came from legitimate offchain sources, with selective disclosure instead of uploading their entire financial lives onchain. lenders could then incorporate those proofs into their risk models. better collateral terms? potentially. personalized lending rates? possible. of course, a verified credit record doesn't eliminate default risk. but forcing everyone to overcollateralize because DeFi can't evaluate anything beyond wallet balances? that's a limitation worth solving. this is where privacy and verification start becoming genuinely useful financial infrastructure.

What if your capital could do more than just support your trades? That’s one of the ideas I find interesting about @grvt_io and GRVT 2.0. With Grvt Earn, the goal is to offer tokenized fixed-income strategies targeting up to 11% APY, while Yield as Margin aims to let users use eligible Earn balances as trading margin and continue earning yield. I like this direction because it focuses on making capital more efficient instead of leaving it idle. Of course, the targeted yield isn’t guaranteed, and the real value will depend on how these features work in practice. Looking forward to seeing GRVT 2.0 bring trading and yield closer together.

The first question I ask about any yield product isn't how much? It's who actually controls my money? A double-digit APY means very little if you don't understand where your funds sit or who has the authority to move them. This is one of the more interesting parts of @grvt_io GRVT uses a self custody model where user assets are held in smart contracts, and withdrawals require authorization from the user's own key. The platform handles trading and execution, while zero knowledge proofs are used to verify settlement on Ethereum. In other words, the exchange infrastructure and the authority over your funds are designed to remain separate. That doesn't mean everything is risk-free. Smart contracts, market exposure and third-party yield strategies still come with their own risks. But I like the direction here. If we're bringing institutional style financial products onchain, user control shouldn't be something we sacrifice to access them.

A $161 trillion market sounds massive, but market size alone doesn't tell the whole story. But the more interesting question is who gets access to that market. For decades, some of the most established fixed-income strategies have been difficult for ordinary investors to reach. High minimum investments, accreditation requirements, limited redemption windows. The infrastructure was largely built around institutions. Tokenization is starting to change that. And @grvt_io wants to take the idea a step further. Its upcoming Grvt 2.0 roadmap combines access to tokenized fixed-income products with something particularly interesting for active traders: using eligible yield-bearing assets as trading margin. The plan is to introduce the upgraded Earn product toward the end of October, followed by yield-bearing Earn balances becoming usable as margin in November. Imagine having capital generate yield while simultaneously supporting your trading positions, without constantly moving it between separate accounts. That wouldn't remove market risk or make returns guaranteed. But it could remove a lot of unnecessary friction between saving, investing and trading. To me, that's where the TradFi onchain narrative starts getting genuinely interesting.

GM web3 💜 Happy Friday!