Screen Girlie 🧑‍💻 (@CryptoJN4) profile picture

Cyber Analyst || Engineer || No code builder || Blockchain Dev || Community builder @LingoAI_io || Building PidginDAO @LingoAI_io

Africa
Joined September 2022
Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)Photo by Screen Girlie 🧑‍💻 (@CryptoJN4)
🎉 PidginDAO Sprint Wraps Up! ⌛Here's a final call to join the PidginDAO sprint on LingoAI’s Zealy questboard! Feel the adrenaline as we close this chapter — but remember, the best is yet to come. zealy.io/cw/lingoai/invite/b…
108
65
4
84
25,826
Screen Girlie 🧑‍💻 retweeted
🏥 Patients can now let an AI assistant read their medical records. athenahealth has launched athenaPatient MCP. Patients who subscribe to Claude Pro or Max in the US can sign in with their patient portal and let Claude pull their lab results, medications, diagnoses, allergies, immunizations and care plans. The controls sound right: patients choose whether to connect, decide what to share, and can revoke access at any time. But athenahealth itself says HIPAA does not apply here. The exchange is patient-directed, and the AI provider is neither a covered entity nor a business associate. Lawyers interviewed agree that in most patient-directed cases, that's correct. So authorising AI to read your records is no longer the hard part. Two questions remain: ① Once the data leaves the hospital, who protects it? ② What exactly is the AI allowed to use it for? A new experimental FHIR implementation guide, Consent About AI, points to an answer for the second question. A patient's choices become rules a system can enforce: ✅ Allow AI for clinical decision support 🚫 Deny use of my data for model training ✅ Allow training only on de-identified data The AI must declare its purpose, and access control enforces the patient's decision. That's the direction we're building toward at @LingoAITech with #Holon Agent 3.0: 🗂 The Holon Pod has data sovereignty, so your health data stays under your control rather than in each AI's session. 📜 Purpose-bound authorisation: consultation yes, model training no, research only if de-identified. 🔐 DID-based identity, so you know which agent and which organisation is asking. ↩️ Revocable authority: withdraw consent and access ends. Health data is some of the most valuable data you own. Sharing it shouldn't mean giving up control over how it's used. #HealthData #AIinHealthcare #DataSovereignty #PatientConsent #FHIR #Holon #LingoAI 👇 (Link to the FHIR guide in the first reply)
133
31
167
2,432
Screen Girlie 🧑‍💻 retweeted
🇸🇬 新加坡金融管理局(MAS)将牵头制定面向金融业的「了解你的智能体」(Know Your Agent,KYA)协议。 金管局局长谢德荣在 #SFF2026 上说得很直白: “有些用户把金融账户的登录凭证交给了智能体。我不建议这样做,安全保障还不够。” 当智能体用客户的账户登录时,银行面临三个问题: ① 客户是否真的授权了这个智能体代为行事? ② 这笔交易是否在智能体的委托范围内? ③ 智能体一旦出错或越权,谁来负责? 金管局的答案不是禁止,而是建立可信协议和安全渠道。KYA 将帮助金融机构核验: ✅ 发出请求的是哪个智能体,它代表谁 ✅ 它被授予了什么权限 ✅ 它按什么委托行事:能做什么、有哪些限额和条件 ✅ 它的来源和可靠性有什么保证 金融机构可以据此决定接受哪些智能体指令,以及按什么条件接受。 正如谢德荣所说:“自主必须有边界,行动及其结果都必须有人负责。” 权限边界的问题已经出现。据路透社报道,Meta 员工在内部反馈称,Muse 提醒一名员工回复 Signal 上的消息,而她表示自己从未授权它访问其他 App。Meta 称报道中的情况属于个别事件。但实际问题依然存在:智能体使用权限的方式,是否符合用户授权时的本意? 🧬 @LingoAITech 正在构建 #Holon Agent 3.0,让个人和组织可以委托 AI 智能体办事,同时掌握自己的数据和决策权。 🔐 基于 DID 的智能体身份,让服务方能验证发出请求的是哪个智能体。 📜 可验证的委托授权,由个人或组织签发,写明智能体代表谁、能做什么、有哪些限制、何时到期。 ↩️ 可撤销的权限,谁授予,谁就可以收回。 🗂 Holon Pod 具备数据主权,由用户决定哪些智能体能访问哪些数据,以及数据可以用于什么目的。 同一套机制可以服务不同场景:家庭可以设定消费限额,企业可以授权智能体向指定供应商付款,公共机构也可以委托智能体执行具体任务。 让智能体比价。要下单,另行授权。要付款,先设额度。 信任不该靠交出钥匙,而应该靠可验证的授权。 #KnowYourAgent #AgenticAI #DID #Singapore #DataSovereignty #Holon #LingoAI businesstimes.com.sg/compani…
64
29
81
1,090
Don't miss the next poker tournament @DustbitOfficial Free to enter, more to win Poker Username: Jjxtz ID: 1-M_kDDC9s
1
39
Screen Girlie 🧑‍💻 retweeted
🚨 $1,000 DUSTBIT POKER FREEROLL! 💰 📅 Thursday, Nov 26th | 1 PM London 600 SEATS. 10 WINNERS. ONE POKER WAR. 🥇 $350 | 🥈 $200 | 🥉 $125 💸 4th–10th share the rest! How to register🧵👇
55
63
11
104
2,964
Send a DM if you have registered 🇹🇭 #BBW2026 #UXUY
🌃 Bangkok nights, great people, and meaningful connections! 🎉 UXUY is bringing the energy to Bangkok with an exclusive Rooftop Blockchain Week Pre-Party ahead of Binance Blockchain Week 2026!
29
Screen Girlie 🧑‍💻 retweeted
🇸🇬 Singapore's central bank will lead the development of a "Know Your Agent" (KYA) protocol for finance. At #SFF2026, MAS Managing Director Chia Der Jiun was blunt: "Some users are giving agents their login credentials to financial accounts. I would not advise giving personal agents such access, as there is insufficient assurance of safety." When an 🤖️ agent uses a customer's login, a bank faces three questions: ① Has the customer really authorised this agent to act? ② Is this transaction within the agent's mandate? ③ Who is accountable if the agent makes a mistake or exceeds its authority? MAS's answer isn't a ban. It's a trusted protocol and a secure channel. KYA would help financial institutions verify: ✅ Which agent is making the request and whom it represents ✅ What authority it has been given ✅ What mandate it operates under: permitted actions, limits and conditions ✅ What assurance exists about its provenance and reliability Financial institutions can then decide which agent instructions to accept, and on what terms. As Chia put it: "Autonomy has to be bounded. Actions and outcomes have to come with accountability." The permission problem is already here. According to Reuters, Meta employees reported internally that Muse reminded one of them to reply to a Signal message, even though she said she had never granted it access to other apps. Meta called the reported cases isolated incidents. The practical question remains: does the agent use its access the way the user intended? 🧬 At @LingoAITech , we're building #Holon Agent 3.0 so people and organisations can delegate to AI agents while keeping control of their data and decisions. 🔐 DID-based agent identity, so a service can verify which agent is making the request. 📜 Verifiable delegation, so a person or organisation can specify whom the agent represents, what it may do, what limits apply and when the mandate expires. ↩️ Revocable authority, so whoever granted the mandate can withdraw it. 🗂 Data sovereignty through the Holon Pod, so users decide which agents can access which data, and for what purpose. The same approach could help a household cap spending, a company delegate payments to approved suppliers, or a public agency authorise an agent to carry out a specific task. Let your agent compare prices. Authorise it separately to place an order. Give it a spending limit before it pays. Trust shouldn't mean handing over your keys. It should mean verifiable authorisation. #KnowYourAgent #AgenticAI #DID #Singapore #DataSovereignty #Holon #LingoAI businesstimes.com.sg/compani…
186
37
1
228
3,482
Screen Girlie 🧑‍💻 retweeted
🌃 Bangkok nights, great people, and meaningful connections! 🎉 UXUY is bringing the energy to Bangkok with an exclusive Rooftop Blockchain Week Pre-Party ahead of Binance Blockchain Week 2026!
2
8
1
10
143
🌃 Bangkok nights, great people, and meaningful connections! 🎉 UXUY is bringing the energy to Bangkok with an exclusive Rooftop Blockchain Week Pre-Party ahead of Binance Blockchain Week 2026!
2
8
1
10
143
Whether you're looking to connect, explore new collaborations, or simply enjoy the company of amazing people in the blockchain space, this is where you want to be! 📅 Date: November 27, 2026 ⏰ Time: 8 PM – 2 AM (GMT+7) 📍 Venue: CHIC Ratchada 18, Bangkok, Thailand
1
1
50
🎟️ Registration is open, but approval is required. Don't miss the chance to be part of the experience. Apply now and secure your invitation! 🔗 Register here 👇 luma.com/pox5so32 See you on the rooftop! 🦄🥂 #UXUY #BinanceBlockchainWeek             #BBW2026 #Bangkok
1
4
70
Screen Girlie 🧑‍💻 retweeted
🚨 $200 DUSTBIT GIVEAWAY!💰 We're giving $100 to TWO lucky people! 💸 Follow @DustbitOfficial Repost & tag 2 friends Drop your DustBit ID below Who's feeling lucky? 👀
1,136
1,030
10
1,114
12,613
If friends are coming to Thailand, you’re welcome to sign up and join us!
UXUY · BBW Bangkok VIP Rides! Whether on-chain or in person, UXUY keeps lowering barriers to participation—making trading simpler and getting together easier. See you in Bangkok, November 27–29, 2026! 💪 This BBW, UXUY is bringing a premium Toyota Alphard fleet to make your Bangkok trip smoother. ✈️ November 27 | Airport Pickups UXUY’s dedicated fleet will offer pickups at Suvarnabhumi Airport (BKK) and Don Mueang Airport (DMK), helping you kick off your BBW trip with ease from the moment you land. 🚐 During BBW | Ride Support Our fleet will be on standby near the venue to help UXUY’s friends and partners get around. Vehicles and ride slots are limited, with rides subject to availability and on-site scheduling. Repost + comment on this post for a better chance of receiving these ride benefits! 🎟️ Sign up now: luma.com/u9lr3mtv
2
3
414
Screen Girlie 🧑‍💻 retweeted
🇸🇬 这周,新加坡给全球 AI 智能体经济出了一道题:当智能体跨平台、跨境行动,信任能跟得上吗? 11 月 17 日,新加坡高级政务部长陈杰豪在 StableX Agentic Network 发布会上说: “挑战不仅是让智能体实现互操作,更是让信任实现互操作,并在智能体行动时持续有效。” 如果每个平台都用一套封闭的身份和授权系统,企业每接入一个新平台,就可能要重新对接。这些成本和复杂度,对中小企业尤其沉重。 次日,副总理颜金勇在新加坡金融科技节介绍了 MAS 与未来金融研究所(FFI)的联合计划。FFI 将设立 Agentic Money Sandbox,让金融机构在受控环境中测试 AI 智能体如何在明确的授权范围和额度内发起支付。 🧬 这正是 @LingoAITech 的 #Holon Agent 3.0 正在构建的信任基础: ✅ 基于 DID 的开放身份。 让智能体使用不依附于单一平台的去中心化标识符,供支持相应标准的系统验证身份。 ✅ Holon Pod。 让用户掌握自己的数据和记忆,并决定谁可以访问。 ✅ 按任务授权,权限可撤销。 比价、下单和付款分别授权。比如,只允许智能体向指定商家支付不超过 100 新元。 ✅ 同一套信任原则适用于不同主体。 无论代表个人、企业、其他组织还是政府机构,智能体都应能证明自己是谁、代表谁,以及被授权做什么。 🌏 让信任成为 AI 智能体经济的基础设施。从新加坡出发,与世界互通。 businesstimes.com.sg/compani… #AgenticAI #DID #Singapore #DataSovereignty #Holon #LingoAI
86
34
103
1,385
GTA 6 has arrived in the UXUY Rewards Center. 🎮 Earn points on eligible spot trades and reserve a digital copy of GTA 6 for 10,000 points. We’re also giving away 5 digital copies of GTA 6, five winners will each receive $100 USD. How to join giveaway Follow @uxuycom & @uxuyzh , then like, reply, and repost this tweet. How to claim reward Winners must claim through a UXUY account that’s more than 3 days old and has completed at least one trade. Giveaway ends November 23. Good luck!
216
215
8
256
11,192
Screen Girlie 🧑‍💻 retweeted
欢迎回来,@JensenHuang 🚀 随着全球 AI 治理格局的迅速碎片化,闭源巨头与开源生态之间、集中式算力与边缘计算之间的张力正变得日益尖锐。@nvidia 不仅是算力的基础设施提供商,更是这场结构性竞争中不可回避的关键变量。而黄仁勋正是这个当仁不让的行业领袖⚙️ 如果 AI 失去控制,或是被少数集中式巨头所垄断,其后果将是不可逆的。⚠️ 开源模型让全球的开发者和公众能够共同审查、监督并改进系统——而不是交出一个黑盒。 边缘计算则将智能下沉至设备与个人,防止形成新型的算力垄断。这正是“以人为本”的核心逻辑:智能应当服务于人,而不是人服务于智能。 🤝 这与 @LingoAITech 以及 #Holon 的愿景高度契合——即创建“主权个人 AI 单元”(Sovereign Personal AI Units),让每个人都能拥有可监督、可控制且能独立运行的 AI。 算力、开源与边缘计算的协同,才是更安全、更可持续的发展路径。✨
Welcome back, @JensenHuang 🚀As the global AI governance landscape rapidly fragments, the tension between closed-source giants and the open-source ecosystem, between centralized compute and edge computing, is becoming increasingly sharp. @nvidia is not only a foundational provider of compute, but also an unavoidable variable in this structural competition. ⚙️ If AI spirals out of control or is monopolized by a few centralized players, the consequences will be irreversible. ⚠️ Open-source models enable developers and the public worldwide to review, supervise, and improve systems together—rather than handing over a black box. Edge computing pushes intelligence down to devices and individuals, preventing new forms of compute monopoly. This is the core logic of putting humans first: intelligence should serve people, not the other way around. 🤝This aligns closely with the vision of @LingoAITech and #Holon—Sovereign Personal AI Units that give everyone AI they can supervise, control, and run independently. The synergy of compute, open source, and edge is the safer and more sustainable path. ✨ #AI #OpenSource #EdgeComputing #AISafety #Holon
65
34
82
928
Screen Girlie 🧑‍💻 retweeted
🤖 本周,全世界都在问同一个问题:AI Agent 到底代表谁? 过去十天里,银行、零售商、监管机构、标准组织以及科技巨头,纷纷在 AI Agent 领域采取动作。 不同新闻背后的核心问题只有一个:当一个 Agent 代表你行动时,凭什么让人信任?它的授权到底由谁赋予、由谁撤销? 🔑 1. 平台与银行正在关上大门 零售巨头反击: 亚马逊、Target 和 Home Depot 正在封杀像 Meta Muse 这类云端购物 Agent。而 Muse 的应对策略是:推出在本地以用户身份运行的浏览器插件。 银行敲响警钟: 在新加坡 🇸🇬,星展银行(DBS)、华侨银行(OCBC)和大华银行(UOB)联合提醒客户防范第三方 AI Agent。OCBC 直接封杀,UOB 隔离运行,DBS 则警告切勿共享账号密码。 👉 核心启示: “把密码交给 Agent”这条路彻底走不通了。身份识别与授权,自始至终都必须掌握在用户自己手中。 ⚖️ 2. 监管机构正在制定规则 🇪🇺 欧盟: 发布《人工智能法案》第50条透明度指南——Agent 必须披露其 AI 身份,并明确说明代表谁行动。 🇬🇧 英国: 四大监管机构(CMA、FCA、ICO、Ofcom)联合发布警告,严防未经授权的数据共享与有偏见的推荐。 🇸🇬 新加坡: 国会辩论了国家级 Agent ID 注册制度,IMDA 正在探索基于 DNS 的身份体系,MAS 则强调事前检查与严格授权。 🏗️ 3. 标准组织正在搭建底层基础设施 🌍 世界银行 & ITU: 世界银行指出数字身份可以确认授权,但委托执行机制仍需从头构建。ITU 本月在巴黎启动了信任与身份(Trust & ID)焦点小组会议。 🌐 Linux 基金会 & W3C: Linux 基金会采纳了 Agent 命名服务(基于 DNS 锚定);W3C 正在就信任层到底采用 DNS 还是 DID 展开辩论。 📜 Sierra & Meta: 发布 Personal Agent Protocol (Poppy) 1.0,支持委托密钥、成对伪名 ID 以及 OAuth 撤销。 🛡️ Plaid、Shufti & Zero Hash: 推出 Know Your Agent (KYA) 工具,强制执行支出限额、有效期控制与即时撤销。 🧩 4. 最关键的一块拼图依然缺失 我们现在有了 Agent 命名(DNS)、协议(Poppy)、身份核验(KYA)以及监管法规(欧盟/英国)。 但依然留下了一个致命追问:授权与记忆究竟存储在哪里?它们到底归谁所有? 如果答案还是“在平台上”,我们不过是把当下的数据垄断 👻 原封不动地带进了 AI 时代。 ✨ 5. Holon 的答案:以用户为起点 Holon Agent 3.0 从第一天设计起,就把用户主权置于核心: 🔐 Solid Pod 授权: 你来决定 Agent 能做什么,并随时撤销权限——无需共享任何密码。 🧠 个人本体(Personal Ontology): 你的 Agent 清晰知道自己代表谁,原生满足欧盟的透明度合规要求。 🎒 可便携记忆:你的上下文始终属于你。随时切换 Agent 或服务,历史记录永不丢失。 🤝 原生开放标准: 完美桥接 DNS 的可发现性与 DID/VC 的可验证性。 🌱 结语 Tim Berners-Lee 曾预言过一个用户通过 Solid 掌握自身数据的 Web 愿景。而 AI Agent 时代的到来,让这一愿景比以往任何时刻都更加紧迫。 @LingoAITech 联合创始人 @HenryWang2021 与 @UnaWCX 始终倡导真正的 Web 3.0 = Data 3.0 + Fin 3.0。在 Agent 经济真正获得信任之前,Agent 的身份、授权与记忆必须重回用户手中。 这正是 Holon 正在构建的事业——与 Inrupt 的愿景携手同行,共筑主权互联网。🚀 #AIAgent #Solid #数据主权 #Holon #LingoAI #Web3
🤖 This week, the whole world asked the same question: who does an AI agent actually represent? In the past ten days, banks, retailers, regulators, standards bodies, and tech giants all moved on AI agents. Different headlines, one underlying question: when an agent acts for you, why should anyone trust it—and who grants and revokes its permissions? 🔑 1. Platforms and banks are closing the door • Retailers push back: Amazon, Target, and Home Depot are blocking cloud shopping agents like Meta's Muse. Muse's workaround? A browser extension that runs locally as the user. • Banks raise red flags: In Singapore 🇸🇬, DBS, OCBC, and UOB warned customers against third-party AI agents. OCBC blocks them, UOB isolates them, and DBS advises against sharing credentials. 👉 The takeaway: "Just give the agent your password" is a dead end. Identity and authorization belonged with the user all along. ⚖️ 2. Regulators are writing the rules • 🇪🇺 EU: Published AI Act Art. 50 transparency guidelines—agents must disclose they are AI and specify on whose behalf they act. • 🇬🇧 UK: Four regulators (CMA, FCA, ICO, Ofcom) jointly warned against unintended data sharing and biased recommendations. • 🇸🇬 Singapore: Parliament debated national agent ID registration, IMDA is exploring DNS-based identity, and MAS stresses pre-action checks and strict permissioning. 🏗️ 3. Standards bodies are building the plumbing • 🌍 World Bank & ITU: World Bank noted digital ID can confirm approvals, but mandate enforcement must be built from scratch. ITU launched a Trust & ID focus group meeting in Paris this month. • 🌐 Linux Foundation & W3C: Linux Foundation adopted Agent Name Service (DNS-anchored); W3C is debating DNS vs. DIDs for the trust layer. • 📜 Sierra & Meta: Released Personal Agent Protocol (Poppy) 1.0 featuring delegated keys, pairwise pseudonymous IDs, and OAuth revocation. • 🛡️ Plaid, Shufti & Zero Hash: Launched Know Your Agent (KYA) tools to enforce spend limits, expiry, and instant revocation. 🧩 4. The most important piece is still missingWe now have agent names (DNS), protocols (Poppy), verification (KYA), and regulations (EU/UK). One critical question remains: Where do authorizations and memories live, and who owns them? If the answer is "on the platform," we are simply importing today's data monopolies👻 into the AI era. ✨ 5. Holon's answer: Start with the user Holon Agent 3.0 was designed from day one with user sovereignty at its core: 🔐 Solid Pod Authorization: You decide what your agent can do and revoke access anytime—no password sharing required. 🧠 Personal Ontology: Your agent knows exactly who it represents, fulfilling EU transparency requirements natively. 🎒 Portable Memory: Your context stays with you. Switch agents or services anytime without losing your history. 🤝 Open-Standards Native: Bridges DNS discoverability with DID/VC verification. 🌱 Closing thought Tim Berners-Lee envisioned a Web where users own their data via Solid. The AI agent era makes that vision urgent. @LingoAITech Cofounders @HenryWang2021 and @UnaWCX advocate real Web 3.0 = Data 3.0 + Fin 3.0. Agent identity, authorization, and memory must return to users before the agent economy can truly be trusted. That’s what Holon is building—walking alongside Inrupt's vision toward a sovereign web. 🚀 #AIAgent #Solid #DataSovereignty #Holon #LingoAI #Web3
61
36
70
958
Screen Girlie 🧑‍💻 retweeted
黄仁勋自诩为一只习惯把头缩在壳里的“乌龟”。🐢 但在开源模型这个议题上,他说这太关键了: “我要把头伸出来,为开源模型据理力争。”但结果你们看到了💥需要有人为他说话,支持他! “开源模型是未来即将诞生的整整一系列产业的基石,我们承担不起开源模型被监管彻底抹杀的后果。” 🌍 —— 黄仁勋@JensenHuang 引自 @nvidia 于 10 月 19 日分享的视频剪辑 有些思想的价值,远胜于任何新闻头条。 无论 @nvidia 未来走向何方,这段话都值得被继续广泛传播。用他原话来说,开源模型“增强了安全与网络安全,加速了创新与传播,并实现了技术主权。它们让每一个开发者、初创公司、大学、行业和国家都能基于 AI 进行构建、定制并从中获益。” 我们还想在此基础上补充一条: 🙋 还有每一个普通人。 🧩 CUDA 将 GPU 转化为人人皆可构建的平台。 🖥️ DGX Spark 将 AI 超级计算机缩小至桌面大小,让大模型在家里就能直接运行。 📡 类似这样的边缘节点,将智能带到了离人们生活更近的地方。 🔐 而下一步,则是要确保这股智能真正听命于它所服务的人。 这正是我们正在与 @LingoAITech Holon Agent 3.0 一起打造的未来 🤖 🧠 单一个人本体(Ontology),协同多个 Agent 智能体 🗂️ 你的数据,完全归属于你自有的主权 Holon Pod 🔁 权限随你掌控,随时可撤销 我们正在将其推向边缘端 📦 LingoBox Nebula——我们自研的硬件设备即将面世:在你拥有的硬件盒子上本地运行模型,数据永远无需离开你的家。🏠🔐 开源模型。边缘智能。属于每个人的技术主权。🌏 领导力问题归属于董事会和正当程序;但开源共荣的精神,属于我们每一个人。🤝 一个本体。多个 Agent。 你的数据。你的智能。你的价值。✨ 🌐 LingoAI.io 🌐 Holon.live #Holon #LingoBox #OpenModels #EdgeAI #DataSovereignty
Entire industries will be built on open models. Developers need the freedom to build them. @JensenHuang on why keeping these models open and accessible matters.
75
38
96
1,909
Screen Girlie 🧑‍💻 retweeted
🤖 This week, the whole world asked the same question: who does an AI agent actually represent? In the past ten days, banks, retailers, regulators, standards bodies, and tech giants all moved on AI agents. Different headlines, one underlying question: when an agent acts for you, why should anyone trust it—and who grants and revokes its permissions? 🔑 1. Platforms and banks are closing the door • Retailers push back: Amazon, Target, and Home Depot are blocking cloud shopping agents like Meta's Muse. Muse's workaround? A browser extension that runs locally as the user. • Banks raise red flags: In Singapore 🇸🇬, DBS, OCBC, and UOB warned customers against third-party AI agents. OCBC blocks them, UOB isolates them, and DBS advises against sharing credentials. 👉 The takeaway: "Just give the agent your password" is a dead end. Identity and authorization belonged with the user all along. ⚖️ 2. Regulators are writing the rules • 🇪🇺 EU: Published AI Act Art. 50 transparency guidelines—agents must disclose they are AI and specify on whose behalf they act. • 🇬🇧 UK: Four regulators (CMA, FCA, ICO, Ofcom) jointly warned against unintended data sharing and biased recommendations. • 🇸🇬 Singapore: Parliament debated national agent ID registration, IMDA is exploring DNS-based identity, and MAS stresses pre-action checks and strict permissioning. 🏗️ 3. Standards bodies are building the plumbing • 🌍 World Bank & ITU: World Bank noted digital ID can confirm approvals, but mandate enforcement must be built from scratch. ITU launched a Trust & ID focus group meeting in Paris this month. • 🌐 Linux Foundation & W3C: Linux Foundation adopted Agent Name Service (DNS-anchored); W3C is debating DNS vs. DIDs for the trust layer. • 📜 Sierra & Meta: Released Personal Agent Protocol (Poppy) 1.0 featuring delegated keys, pairwise pseudonymous IDs, and OAuth revocation. • 🛡️ Plaid, Shufti & Zero Hash: Launched Know Your Agent (KYA) tools to enforce spend limits, expiry, and instant revocation. 🧩 4. The most important piece is still missingWe now have agent names (DNS), protocols (Poppy), verification (KYA), and regulations (EU/UK). One critical question remains: Where do authorizations and memories live, and who owns them? If the answer is "on the platform," we are simply importing today's data monopolies👻 into the AI era. ✨ 5. Holon's answer: Start with the user Holon Agent 3.0 was designed from day one with user sovereignty at its core: 🔐 Solid Pod Authorization: You decide what your agent can do and revoke access anytime—no password sharing required. 🧠 Personal Ontology: Your agent knows exactly who it represents, fulfilling EU transparency requirements natively. 🎒 Portable Memory: Your context stays with you. Switch agents or services anytime without losing your history. 🤝 Open-Standards Native: Bridges DNS discoverability with DID/VC verification. 🌱 Closing thought Tim Berners-Lee envisioned a Web where users own their data via Solid. The AI agent era makes that vision urgent. @LingoAITech Cofounders @HenryWang2021 and @UnaWCX advocate real Web 3.0 = Data 3.0 + Fin 3.0. Agent identity, authorization, and memory must return to users before the agent economy can truly be trusted. That’s what Holon is building—walking alongside Inrupt's vision toward a sovereign web. 🚀 #AIAgent #Solid #DataSovereignty #Holon #LingoAI #Web3
150
39
2
179
5,226