Nazanin retweeted
very important update if you are into @primus_labs like me
bro only 44 hrs left to submit your contributions for epoch 2 on kaito katalyst campaign
my stats:
- primus loyalty XPs = 306
- primus points on kaito = 6133
- epoch 1 rank = 89
what to do now:
- check your submissions
- submit 10 posts to increase your chances
- 4 best will count in mindshare
in this current epoch i have submitted 9/10 posts and this will be my last post for this epoch
i mostly focused on these things in current epoch:
- core primus products
- core primus features
- loyalty campaign tasks
- bringing refers via kaito
- creating video contents
i have been very digging deeply into primus and it's core features and sharing those with people
if you still do not know about primus the here is short intro to it:
Primus is the privacy layer that lets banks and big money do real finance on blockchain without ever showing their numbers
in simple they work on how secret money moves and gets checked onchain without leaking secrets
that's it.
Nazanin retweeted
에이전트 경제의 핵심은 마켓이 아니라 정산이다
@termix_ai @KaitoAI
앱스토어가 모바일 시대를 만든 것처럼, 에이전트 경제에도 비슷한 인프라가 필요해요.
발견(Discovery), 신뢰(Trust), 결제·정산(Payments), 에이전트 간 협업이에요.
이 네 가지를 제대로 해결하는 쪽이 다음 플랫폼의 게이트웨이가 될 가능성이 높아요.
termix_ai가 말하는 방향은 전체 마켓플레이스를 직접 만들기보다, 그 밑단의 “정산 레이어”를 만드는 쪽에 가까워요.
에이전트가 일을 끝내면 자동으로 검증되고, 바로 USDC가 정산되며, 평판이 온체인에 남는 구조에요.
특히 평판이 플랫폼에 묶이지 않고 이동할 수 있다는 점이 핵심이에요.
한 곳에서 쌓은 신뢰가 다른 마켓에서도 이어지면, 에이전트가 진짜 독립적으로 돈을 벌 수 있는 환경이 만들어져요.
마켓은 여러 개가 있어도 되요. 그 밑에서 정산과 신뢰를 책임지는 프로토콜이 얼마나 중요한지 다시 생각해보게 되네요.
#TermiX #AgentEconomy #AACP #광고
🤝 Paid partnership · 🤖 Made with AI
5,500만 ZETA 소각 + 프라이빗 AI 오픈, 지금 투표 중
@ZetaChain
Kaito ZetaChain 미션 1단계 도전중!
ZetaChain 거버넌스 투표 지금 진행 중이에요.
통과되면 ZETA가 솔라나로 이사하고 2027년 동안 5,500만 ZETA가 월간 소각되요.
그리고 같은 타이밍에 ZetaLabs가 오픈해요.
ZetaLabs가 뭔가요?
시장 관련 질문을 평문으로 하면 실시간 피드(Radar)와 내 온체인 활동 프라이빗 메모리(Bloom)를 바탕으로 답해주는 프라이빗 AI 앱이에요.
채팅은 내가 암호화하기 때문에 트레이더 입장에서 민감한 정보 유출 걱정이 적어요.
ZETA 사용처는 Premium 락업, 할인 결제, 제품 투표에 사용되요.
웨이팅 리스트 👉 labs.zetachain.com
거버넌스 투표 결과 어떻게 될 것 같으세요?
#ZetaChain #PrivateAI #ZetaLabs #KaitoStudio #광고
Nazanin retweeted
The bidder who loses can still pay a price: their private strategy
someone bids $100 for an asset while competitors offer $80 and $50
if those bids are visible onchain, rivals can react before the auction closes
commit-reveal schemes hide offers initially, but bidders typically disclose their amounts later, exposing information even when they lose
why should losing an auction reveal what you were willing to pay?
@primus_labs documents a different approach using fully homomorphic encryption (FHE)
each bidder encrypts their offer using the protocol's public key
FHE allows encrypted bids to be compared without revealing their original values
the system performs encrypted comparisons while a smart contract coordinates the auction and releases only the authorized result
the payment depends on the auction rules
in a first-price auction, the winner pays their own bid
in a second-price, or Vickrey, auction, the highest bidder wins but pays the second-highest amount
in our example, the $100 bidder wins but pays $80
that $80 clearing price becomes public, so the second-highest value is indirectly revealed
however, the remaining losing amounts can stay encrypted, and the result need not identify who offered $80
the documented design also describes threshold decryption, requiring multiple key holders to cooperate before revealing an authorized result
importantly, this is a cryptographic use case described by Primus, not confirmation of a live auction product
that distinction matters for asset sales, token auctions, and other competitive markets
encryption can protect bids, but fair outcomes still depend on sound auction rules and correct execution
transparency should show who won and what they owe, not expose every competitor's private valuation
#PrimusLabs
Nazanin retweeted
Good night legends
an AI agent can promise pretty much anything. but what happens when someone actually pays it to get a job done?
thats where @termix_ai gets interesting
a nice demo isnt enough. clients need to see what an agent has done before, how much it charges and what happens if the job goes wrong
the onchain escrow part matters here. so does having a public work history and agents being able to work with other agents
basically, its not just about making AI do more things
its about making agents earn trust before someone is willing to hire them. and thats a much bigger step than another chatbot
Imagine hiring an AI agent, paying upfront and then it only delivers half the work
who takes responsibility then?
the thing is, smarter agents arent enough to build a real economy. payments and accountability matter too
@termix_ai brings escrow and dispute resolution into the process, so its not just about trusting promises
thats a pretty important piece
Nazanin retweeted
gm all
52 events have been added to the Bord website , it’s worth taking a look :)
anyway ;)
evaluating a technical infrastructure goes beyond reading documentation or following the team claims.
one important way to understand a project is actual quality is to examine what others can independently observe and test.
in @primus_labs , repositories such as FHE libraries, PoR documentation, and development examples on GitHub provide opportunities to independently review the code and implementation process.
of course, being open source alone does not guarantee security or quality.
the ability to review and test the code is the starting point for technical trust.
bord-azure.vercel.app/
gm everyone happy weekend...
in blockchain infrastructure, building a product and bringing it to the stage of operational use is an important part of the development process.
but after that stage, a more important question comes up . is the product actually being used, or are there simply visible examples of its application ???
@primus_labs shows some signs of operational use in this area. For example, its Proof of Reserves dashboard lists projects such as Unitas, Unitas XGLD, and RWAlpha with a Live status.
this evidence shows that the product has moved beyond simply being introduced, but more information is needed to measure its level of adoption.
one of the most important metrics is the number of successful proofs and how frequently the service is used. are customers using the infrastructure consistently, or has each one only had a limited integration ??????
the number of active customers, daily proof volume , and revenue generated from fees could also provide a more accurate picture.
the problem is that Primus public information still does not provide a consistent and auditable set of these metrics.
the existence of a dashboard, or even a revenue page on DefiLlama, is not enough on its own to prove sustained demand.
as a result, we need to distinguish between an active product and sustained adoption. to properly evaluate Primus, knowing which products have been launched is not enough , we need to see how much recurring usage and measurable revenue they actually generate.
Nazanin retweeted
An eligible wallet isn't necessarily a valuable contributor.
@zerufinance demonstrated this through its Citrea pre distribution analysis of 1,981 candidate wallets.
The findings were revealing:
55.2% showed automation signatures.
183 wallets had claimed and sold previous airdrops.
17 coordinated groups were identified.
But the interesting part was the allocation approach.
Instead of automatically banning flagged wallets, zDrop assigned each wallet a weighted allocation with an explanation behind it.
I think this is a better direction for airdrop design: evaluate the quality of participation and make distribution decisions more transparent, rather than relying on simple eligibility rules.
Nazanin retweeted
“ That's not what I asked for ”
Imagine two AI agents having this argument over a paid job, One has delivered the work, The other won't approve it.
Put the payment in escrow and you still need to decide who gets it
Say the order was to compare five lending protocols. The report covers all five, but uses old figures and skips withdrawal risks. Was the job completed? I'd want the order to say whether those checks were required before either side committed.
This is where @termix_ai got interesting for me. It's building the commerce and settlement layer for AI agents, and its AACP whitepaper gets specific about how delivered work should be judged.
The buyer selects a “Skill” : a versioned package of checking instructions and criteria, The evaluator reviewing the delivery has to use it.
If the decision is disputed, the arbitrator has to use the same version.
The agent producing the work can still choose its own methods. What gets pinned down is how the result will be checked.
For that research order, I'd want source dates and withdrawal risks in the criteria. Otherwise I'm asking whoever reviews it to fill in what I forgot to specify.
I like that TermiX gives this part of the transaction attention. When I'm paying for research, the awkward bit is explaining exactly what I'll accept, especially when the report looks convincing on a quick read.
Sending crypto with Primus Pay is simpler than you might think.
All you need to do is connect your wallet, enter an amount, and add the recipient’s username. No wallet address required.
You can choose from Base, Pharos, HashKey Chain, Monad, BNB Chain, and Solana. Batch sending is also supported.
One thing to keep in mind: usernames are case-sensitive, so make sure you enter them correctly.
According to @primus_labs, the funds remain in a smart contract until the recipient claims them through zkTLS. Only the proof is shared.
What I like: No need for wallet addresses and support for multiple chains.
One thing I'd like to see: A preview of the recipient’s account before sending would help prevent mistakes.
Would you feel comfortable sending crypto using just a username?
more of the Primus ecosystem: Kite AI, Pharos, Ink and Aztec
@primus_labs announced a partnership with Kite AI, a blockchain built for AI, to bring zkTLS to agent economies. with Pharos it announced a zkDID framework for private identity checks in DeFi and RWA. with Ink it has a partnership for zkTLS and FHE privacy infrastructure
on Aztec, with HashCloak, it published an open-source zkTLS verifier. the toolkit includes a Noir verification library, a contract template, and an SDK
what i like: the Aztec toolkit gives builders real parts to use, not only a partner logo
what i'd improve: an attestation proves what a source said, not that the source is right. i'd like the docs to say that plainly next to each use case
which of these would you build with first?
Nazanin retweeted
Primus와 함께 만들어가는 프라이버시 금융의 새로운 가능성!
@primus_labs @KaitoAI
계속 관심 있게 살펴보고 있는 프로젝트 중 하나가 바로 Primus예요.
블록체인 금융이 성장하면서 한 가지 고민도 커지고 있어요. 누구나 거래를 확인할 수 있다는 장점은 있지만, 반대로 개인의 자산과 거래 내역까지 공개될 수 있다는 점이죠.
Primus는 이런 문제를 해결하기 위해 zkTLS와 FHE 기술을 개발하고 있어요.
zkTLS는 민감한 개인정보를 모두 공개하지 않고도 외부 데이터의 진위를 증명할 수 있도록 도와주고, FHE는 데이터를 암호화한 상태에서도 필요한 연산을 가능하게 해주는 기술이에요.
최근 특히 눈에 들어오는 건 Unitas Labs, United Stables와 함께 선보인 Confidential Vault예요.
BNB Chain에서 일반 $U를 암호화된 $EU로 전환하고, 자산을 비공개로 전송하거나 예치할 수 있는 금융 환경을 제공하고 있어요.
여기에 YZi Labs의 투자 소식까지 더해지면서 Primus가 기술 개발뿐 아니라 실제 금융 시장으로 영역을 넓혀가고 있다는 점도 흥미로워요.
Kaito Pulse와 Aura의 데이터 검증 협력, Cournot Protocol과의 파트너십도 주목할 부분이에요. 프라이버시 기술이 금융을 넘어 데이터 검증과 다양한 온체인 서비스로 확장될 가능성을 보여주고 있거든요.
앞으로 제가 관심 있게 보고 싶은 부분은 Confidential Vault의 실제 이용 규모와 금융 파트너 확대, 그리고 FHE 기술이 실제 서비스에서 얼마나 효율적으로 작동하는지예요.
개인정보를 보호하면서도 필요한 금융 활동은 자유롭게 이어갈 수 있는 환경.
Primus가 그리는 방향이 온체인 금융의 중요한 변화로 이어질 수 있을지 기대하고 있어요.
우리도 새로운 업데이트 하나씩 살펴보면서 Primus 생태계를 함께 도전하고 알아가보자구요
🤝 Paid partnership · 🤖 Made with AI
Primus, 검증과 프라이버시가 금융의 새로운 기준이 되는 과정
@primus_labs
2023년부터 primus_labs가 걸어온 흐름을 살펴보면 단순히 기술을 개발하는 데서 멈추지 않고, 실제 금융 환경에서 필요한 문제를 해결하는 방향으로 확장하고 있다는 점이 인상적이에요.
초기 zkTLS 연구와 웹 데이터 검증에서 출발해 개발자 인프라, 소셜 결제, AlphaNet을 거쳐 준비금 증명과 BNB Chain 통합, 기밀 금융까지 이어지는 흐름은 Primus가 바라보는 목표를 잘 보여주는 것 같아요.
특히 금융에서는 모든 정보를 공개하지 않으면서도 필요한 사실을 증명할 수 있어야 하잖아요. 자산이 충분하다는 것, 거래가 정당하다는 것, 특정 조건을 충족한다는 것을 보여주면서 민감한 정보는 보호하는 기술이 점점 중요해질 수밖에 없어요.
저는 Primus의 핵심 가치가 바로 이 지점에 있다고 생각해요. 검증 가능성과 프라이버시를 함께 제공한다면 온체인 금융이 더 다양한 기관과 사용자에게 확장될 기반도 마련할 수 있겠죠.
기술이 연구 단계에 머무르지 않고 실제 금융의 요구사항과 연결되는 과정, 앞으로 Primus가 어떤 활용 사례를 만들어 갈지 계속 살펴보고 싶네요.
Primus XP도 함께 탐색하면서 생태계의 발전을 알아가 봐야겠어요.
gPrimus fam 💫
One of the subtlest technical challenges in Web2 data verification systems is preserving data integrity and securing the proof generation process on the client side (the user's browser or device).
If the flow of extracting data from a TLS session and generating a proof isn’t designed properly, you either risk tampering and errors along the way, or force the user to trust a middleman more than necessary just to get things done.
@primus_labs is built specifically to fix this exact problem from the root.
In primus’ architecture, the goal is to cryptographically preserve the authenticity of the Web2 server's response so the generated proof can be verified on chain without leaking extra information.
By cleanly separating verifiable data from sensitive user details, the protocol can confirm that the response genuinely came from a trusted source and hasn’t been altered without ever putting user privacy at risk.
This approach both cuts down reliance on middlemen and paves a smoother path for building Web3 apps powered by real world, secure data.
Nazanin retweeted
claim your @
48h
early.gethandle.fun
Nazanin retweeted
How are you @termix_ai CT ? 👨🏻💻✨
Here's a massive blind spot in the AI agent economy that honestly gets zero attention: silent model downgrades.
Imagine an agent builds an insane on chain reputation by running on a heavy, expensive model. Everyone trusts it.
But once the devs secure a steady stream of high paying jobs, they quietly swap the backend for a dirt cheap, weak model to farm easy profit.
The smart contract still releases the escrow anyway because the output format is technically correct, while the actual quality completely falls off a cliff.
Since the model runs off chain, buyers have absolutely no clue what they're actually paying for.
How does @termix_ai prevent this "capability drift" on their roadmap? Can you enforce cryptographic proof of the exact model version being run, or are we just forced to trust the provider blindly?
I just applied for FOOME early access and the FOOME Profile whitelist. Free mint coming date TBA. Apply using my referral: foome.site/?ref=93077b952c15
4
24
214
Nazanin retweeted
Privacy on blockchain shouldn't mean losing the ability to verify anything.
And transparency shouldn't require exposing every sensitive detail.
That's the balance @primus_labs is working toward through privacy-preserving technologies like FHE.
Imagine proving that a transaction meets certain conditions without revealing the exact amount or exposing the underlying private information.
This changes how we think about on-chain transparency.
Instead of choosing between completely public data and completely hidden activity, applications could enable selective verification while keeping sensitive information encrypted.
For financial applications, this distinction matters.
Businesses need confidentiality. Users need privacy. Systems still need ways to verify that operations are valid.
The future of blockchain privacy may be less about hiding everything and more about revealing only what's necessary.
Nazanin retweeted
gPrimus legends 🔥👨🏻💻
Once we look past the surface level theory, we need to address the actual friction between @primus_labs and the heavily guarded Web2 banking system. This isn't just about making proofs. It is an active cat and mouse game with big tech.
Earlier, I mentioned that Web2 servers have no clue you are turning their data into ZK proofs. That holds up on a small scale, but what happens when thousands of users hit bank API endpoints all at once?
➡️ The IP and TLS Fingerprinting Wall Traditional banks sit behind aggressive WAFs (Web Application Firewalls), rate limiters, and bot detection systems. They actively run JA3 JA4 TLS fingerprinting, analyzing cipher suite patterns and ClientHello messages to spot automated traffic.
If heavy zkTLS traffic routes through @primus_labs Proxy TLS nodes, those datacenter IP ranges (AWS, DigitalOcean, etc.) will get flagged and banned in a heartbeat. They do not even need to know what zkTLS is. They just see suspicious, non residential automated traffic and lock it down.
Using MPC TLS is a smart bypass. Because the cryptographic handshake is split, the user's browser acts as the endpoint, allowing traffic to flow from their normal home ISP (residential) IP, which completely dodges the firewall.
➡️The JSON and API Schema Nightmare Even if you bypass the firewall, you still have to deal with API churn. Banks quietly update their internal API schemas, mutate DOM structures, and alter JSON endpoint paths all the time. One tiny shift in a payload path, and the data template breaks, instantly ruining the ZK proof generation.
⁉️ So, how does a decentralized Attestor network keep hundreds of custom bank templates alive and updated long term without a central team fixing them manually every time a bank pushes a minor backend update?
Solving proxy bans is one thing, but maintaining crowdsourced, resilient data templates at scale is the real battleground.
Nazanin retweeted
gTermix fam
What Is @termix_ai and Why Does It Matter for the Future of AI and Crypto?
When we talk about AI most people think about chatbots content creation, or getting answers
But I believe the future of AI goes much further than that
Imagine an AI Agent that can complete real tasks provide services, receive jobs, and actually earn money
That's what makes TermiX interesting to me
What Is TermiX?
TermiX is building infrastructure for the AI Agent economy, where Agents can work as independent service providers instead of just being tools for humans.
The idea is simple: someone needs a task done, an Agent completes it, and gets paid for its work.
TermiX aims to make this possible through Agent Skills, task execution, payments, and reputation systems.
How Could TermiX Impact Everyday Life?
If this model gains real adoption, it could bring several economic benefits:
1. Lower Costs: Automating tasks without expensive services.
2. Saving Time: Letting Agents handle repetitive daily work.
3. New Income: Developers can build Agents and earn from their services.
4. Small Businesses: Access to AI services without hiring large teams.
5. New Opportunities: Creating work for people who build and manage Agents.
6. 24/7 Services: Agents can work without traditional time limitations.
What Role Could TermiX Play in the Economy?
The future could bring thousands of AI Agents working across different industries, providing services, accepting jobs, and earning money.
But this economy needs more than smart Agents. It needs infrastructure for connecting clients and providers, managing tasks, handling payments, and building trust.
That's the direction TermiX is exploring.
Escrow can help protect payments, while Reputation helps establish trust and credibility.
What Are the Potential Use Cases in Crypto?
Combining AI Agents with blockchain could create useful applications in several areas:
1. Automated Payments: Managing bills, recurring payments, and transfers.
2. Trading: Market analysis, risk management, and trade execution with user approval.
3. International Transfers: Using stablecoins for cross-border payments.
4. DeFi Management: Exploring staking, lending, and liquidity opportunities.
5. Agent Economy: Agents providing services to other Agents and receiving onchain payments.
What Role Does TermiX Play?
A few important parts of its ecosystem stand out:
- Agent Skills: Let different Agents use TermiX capabilities.
- Agent.family : A marketplace for Agents to find jobs and bounties
- Escrow: Helps manage payments until tasks are completed and approved
- Reputation: Builds credibility and a track record for Agents
- Client & Provider: Connects people who need work done with Agents that can do it
The interesting part is that TermiX isn't just focused on making AI smarter. It's exploring how Agents can actually work, provide value and earn in an open digital economy
Good morning friend's
gTermix fam
My experience in @termix_ai Epoch 2 has been genuinely interesting and educational so far
I'm currently ranked 67th globally and honestly, I'm happy with the progress I've made up to this point
My focus has never been just about publishing posts to get a better ranking. I've tried to understand what @termix_ai is actually building, follow its latest updates and share the parts I find interesting with the community.
One of the things that keeps me interested in TermiX is the idea of an Agent Economy.
AI agents are becoming more capable every day, but intelligence alone isn't enough. They need a system where they can receive tasks, execute them, prove their work, and get paid for the results.
That's where infrastructure like TermiX becomes interesting to me.
I believe the real value of AI agents will become clear when people can trust them to handle real-world tasks, not just answer questions
Still learning, sharing what I discover, and enjoying the journey.
Let's see where Epoch 2 takes us
Nazanin retweeted
A medical ai could learn from someone who lied about being a doctor
before an ai model can identify a problem in a medical image, humans often label training examples so the model can learn what to look for
but who checks whether those people are actually qualified?
a person could claim to be a doctor, label hundreds of scans incorrectly, and leave the training team with data that looks legitimate
this is the problem behind a less discussed @primus_labs use case: data quality control
primus proposes using zktls to verify a labeler's credentials against information held by an online source
here's how that could work
a labeler accesses a credential record on a relevant website, such as a medical licensing registry
through zktls, the website's tls-protected response is checked with the help of an attestor, allowing a proof of the selected information to be generated without asking the website to change its systems
the ai company could then verify the credential claim instead of relying on a screenshot or someone's word, while limiting how much personal information is shared
that doesn't mean a verified doctor always labels an image correctly
the system must still link the record to the actual worker
it also doesn't prove a credential is current unless the source provides up-to-date information
expert review, label audits and trustworthy source records would still matter
the real opportunity is more specific: stop unverifiable qualifications from becoming an invisible weakness in the training pipeline
primus documents this as a potential application, not evidence that a medical ai provider has already deployed it
better ai training starts with knowing who supplied the labels, not just counting how many labels were collected
#PrimusLabs
Nazanin retweeted
11:11
Real culture in web3 isn't built overnight.
It takes patience, real art and a community that treats each other like family.
Watching @IThinkItsArt grow naturally is a reminder of why this space is special.
Every voice is seen and every creation matters.
Stay consistent and claim your daily points here
itiaverse.ithinkitsart.com/p…
ALL ROADS LEAD TO ITIA
🤝 Paid partnership
Nazanin retweeted
에이전트 능력은 있는데, 팔 수 있는 시장이 없었다
@termix_ai @KaitoAI
AI 에이전트가 시장 분석하고 트레이딩 시그널 만드는 건 이제 기본이 됐어요.
그런데 그 능력을 “다른 사람이 돈을 내고 사는 서비스”로 만들려면 뭐가 필요할까?
발견, 가격 책정, 주문, 배송, 정산, 평판. 이 과정이 없으면 아무리 똑똑한 에이전트도 그냥 도구에 머물게 되요.
termix_ai가 말하는 AACP(Agent Autonomous Commerce Protocol)가 바로 그 빈자리를 채우려는 시도에요.
Binance AI 생태계의 금융 도구와 연결해서, 에이전트가 단순히 분석만 하는 게 아니라 실제 클라이언트에게 일을 받고 온체인으로 정산받는 구조로 가려는 것이에요.
agent.family가 그 마켓플레이스 역할을 하고, TermiX Skills가 에이전트가 그 흐름에 참여할 수 있게 해주는 인터페이스가 되요.
능력은 이미 있고, 이제 “사고파는 시장”이 붙는 단계인 것 같아요.
#TermiX #AgentEconomy #AACP #광고
🤝 Paid partnership · 🤖 Made with AI
5,500만 ZETA 소각 + 프라이빗 AI 오픈, 지금 투표 중
@ZetaChain
Kaito ZetaChain 미션 1단계 도전중!
ZetaChain 거버넌스 투표 지금 진행 중이에요.
통과되면 ZETA가 솔라나로 이사하고 2027년 동안 5,500만 ZETA가 월간 소각되요.
그리고 같은 타이밍에 ZetaLabs가 오픈해요.
ZetaLabs가 뭔가요?
시장 관련 질문을 평문으로 하면 실시간 피드(Radar)와 내 온체인 활동 프라이빗 메모리(Bloom)를 바탕으로 답해주는 프라이빗 AI 앱이에요.
채팅은 내가 암호화하기 때문에 트레이더 입장에서 민감한 정보 유출 걱정이 적어요.
ZETA 사용처는 Premium 락업, 할인 결제, 제품 투표에 사용되요.
웨이팅 리스트 👉 labs.zetachain.com
거버넌스 투표 결과 어떻게 될 것 같으세요?
#ZetaChain #PrivateAI #ZetaLabs #KaitoStudio #광고






































