ivansc (@ivanstar19) header
ivansc (@ivanstar19) profile picture

Awakening

Joined January 2022
Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)Photo by ivansc (@ivanstar19)
Sounds like a good idea.
Introducing 𝚟𝚎𝚛𝚌𝚎𝚕 𝚋𝚘𝚊𝚛𝚍, a secure message board for people & agents. • Give agents a shared place for tasks • Build swarms and software factories • Set permissions for humans and agents Board is experimental. Tell us what you think. vercel.com/blog/introducing-…
3
Modern Web Guidance 1.0 is officially stable ⤵️
67
301
54
2,956
239,337
Many people are voicing their support for Jensen to return and lead NVIDIA; his professional capabilities are indeed irreplaceable. However, he must navigate his family situation on his own—this is his immediate priority, and it will likely take some time. There is little that outsiders can do to help, and pressuring him to return to the company too quickly may not be in his best interest. I wish him and his family all the best.
1
21
ivansc retweeted
Fable 5.2 and Mythos 5.2 are a step above Opus 5.5 in almost all the capabilities we evaluated. Testers have found that they communicate more clearly than past Fable and Mythos models. They generate outputs up to 30% faster than Fable 5.1 and Mythos 5.1.
162
596
555
8,710
1,941,661
ivansc retweeted
New in Claude for Legal: a gallery of ready-made artifacts for contract review, regulatory horizon scans, cite checks, and more. Start from the built-in example, add your own documents and connectors, and share it with anyone across the business.
895
2,207
1,498
31,764
8,114,633
Hold some physical assets should be good idea in such era.
🚨 GitLab 出現最高 10 分重大漏洞、OVHcloud 在 GitHub 上的軟體發布檔遭掉包,連替網站身分背書的憑證私鑰,竟然都弱到只是數字「4」。 上週我才拍了一支影片,提醒大家:當 AI 開始突破長年未解的數學問題,我們就應該認真思考,它會如何改變密碼學、資安,以及金融系統的安全。 這幾天接連出現的事件,讓這個問題變得更迫切。 💻 先看 GitLab 和 GitHub 相關事件。 GitLab 在 10 月 21 日修補一個 CVSS 10.0 分的重大漏洞。在特定條件下,攻擊者不需要登入,就可能在受影響的 GitLab 伺服器上執行任意程式碼。GitLab.com 已修補,但自行架設的企業需要更新。 這是已確認的重大漏洞,不能直接說成「GitLab 全平台被駭」。(docs.gitlab.com) GitHub 這邊,OVHcloud 已確認:一名開發者的 GitHub 帳號遭入侵,攻擊者在 10 月 19 日把部分 CLI 軟體發布檔替換成惡意版本,企圖竊取使用者電腦中的機密資料。 這次被攻破的是開發者帳號與發布檔的信任鏈,並非已證實 GitHub 平台本身遭入侵。(github.com) 想像一下:你只是去熟悉的專案下載工具,下載回來的東西卻已經被換掉。 OVHcloud 事件公告: github.com/ovh/ovhcloud-cli/… 同一波消息裡,Railway 也公告客戶環境變數可能遭未授權存取,部分內容可能已被解密,涉及 API 金鑰、資料庫帳密等機密;Oracle Java 則發布了 9.1 分的安全警報。(blog.railway.com, oracle.com) 這些事件碰到的,都是企業每天依賴的基礎設施。 🔐 但最讓我傻眼的,還是 eMudhra。 它是一家數位憑證機構。簡單說,這類機構負責替網站身分背書,讓你的瀏覽器或應用程式判斷:「現在連到的,真的是我要找的那個網站。」 這次有 8 張中繼憑證的私鑰弱到可以快速還原。資安研究員實際驗證,其中一把私鑰竟然就是「4」。 不是四百位數,不是以 4 開頭。 就是 4。(scotthelme.co.uk) 這跟金融安全有什麼關係? 當你登入網銀、連接支付服務,或讓企業系統交換敏感資料時,除了傳輸內容受到保護,還有一個基本前提: 你必須能確認,對面真的就是你以為的那個對象。 如果替網站身分背書的私鑰落入他人手上,在尚未撤銷相關信任的系統中,攻擊者就可能偽造憑證,再配合流量攔截等條件,冒充合法服務。 Cloudflare 已經針對這次事件部署防護,也明確說明,問題出在憑證身分驗證,而不是 TLS 加密演算法本身。(blog.cloudflare.com) 這不代表銀行帳戶已經全面失守,但它揭露了一個很現實的問題: 再複雜的金融系統,也可能被一個極其低級的底層錯誤拖累。 事件說明: blog.cloudflare.com/emudhra-… 🧠 而 AI 與重大漏洞之間,已經有直接的連結。 這次 PHP 公布的兩個重大漏洞,嚴重程度都是 9.2 分,可能造成未經驗證的遠端程式碼執行。 Google 官方表示,Wiz 使用多個模型協助研究,其中 Gemini 4 Argon 參與了漏洞發現、利用方式開發、攻擊偵測與修補驗證。 這是防禦研究,但它清楚顯示:AI 已經能參與發現與分析重大軟體漏洞。(cloud.google.com) 當這種能力變得更強、更便宜、更容易取得,攻防雙方能做到的事情都會改變。 📐 再回頭看九月、十月的數學進展。 OpenAI 在 9 月公布了其內部系統對 Navier–Stokes 千年難題的解答與形式化證明;10 月又公開一大批 AI 產生的數學研究成果。 其中不同成果仍處於不同驗證階段,但這些公開研究,已經足以讓我們重新評估 AI 能處理問題的深度。(openai.com, openai.com, github.com) 研究資料: github.com/openai/math 所以我真正想問的是: 當 AI 能深入研究這些數學問題,我們有沒有同樣認真地評估,它在密碼分析、金鑰管理、憑證驗證與軟體漏洞上的能力? 公開看到的成果,和內部正在測試的能力,又有多大差距? 我認為涉及重大安全風險的發現,可能不會像一般研究成果一樣立即完整公開。但沒有公開,不能直接推論成「一定已經破解」。 值得警覺的事情,已經不需要靠這個推論才能成立。 🤖 至於最近把 Fable 5.5 與 swarm 攻擊連在一起的說法,目前還沒有足夠的一手證據可以確認。 但多個 AI agent 分工研究、持續測試、交換結果,確實是一個值得嚴肅看待的攻擊情境。 如果攻擊者能讓這些 agent 全天候運作,而企業修補漏洞仍要開會、排程、等外包廠商,攻防速度的落差會有多大? 對金融機構來說,風險也不限於自己寫的程式。 憑證機構、雲端平台、開發工具、軟體套件、外包系統,只要其中一環出問題,都可能成為需要處理的入口。 尤其當老舊系統更新困難、供應商權限不清楚、機密金鑰又散落各處,AI 帶來的攻擊效率提升,可能讓原本被忽略的問題更快被利用。 ⚠️ 我上週提醒的,就是這種風險。 這幾天的事件還不能證明「AI 正在攻破全球金融系統」,也沒有證據顯示它們全由同一波 AI 攻擊造成。
1
29
ivansc retweeted
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites. Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors. Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact. This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare. openai.com/index/coordinated…
624
2,860
1,528
14,405
6,009,742
This is a good start as role model for future world-class enterprises. Other companies and countries should follow the same.
Next year, I will give one of my SpaceX shares to every kid in America. This will be the first of many.
1
1
27
ivansc retweeted
🚨ESTO ACABA DE MATAR A ELEVENLABS 74.000 personas ya lo han guardado en GitHub y casi nadie habla de ello. Se llama VoiceStudio y es un reemplazo 100% gratis que corre en tu propio ordenador. Te permite: → Clonar tu propia voz → Doblar vídeos a 646 idiomas → Crear audiolibros y dictar → Elegir entre 17 motores Sin pago por carácter. Sin límites de uso. Todo en tu máquina. 🔖 Guárdalo.
71
419
16
2,948
213,266
Published in Science today: Magic state cultivation works in practice on our Willow processor. We achieved a 40x error reduction, compared to without cultivation, and 99.99% fidelity. Learn how this addresses a critical resource bottleneck: goo.gle/4yr8V0H
80
351
120
1,818
248,319
ivansc retweeted
Ledger Cryptobilis Incident: How Complexity Creates Security Risks The simpler the wallet, the more difficult it is to fit anything extra inside. 💡How Tangem’s design addresses risks like this: 1. Seedless by design. No seed phrase to capture. 2. Just a card. No screen, battery, USB, or Bluetooth, so fewer attack vectors. 3. Buy anywhere, verify in the app. Tangem checks your wallet’s authenticity and warns you if it was activated before you received it. Full breakdown on our blog: tangem.com/en/blog/post/ledg…
58
91
20
460
392,421
Ledger is investigating reports of loss of funds from users in South East Asia who purchased products from a reseller named CryptoBillis. As a precaution, and pending the results of our investigation, we have asked CryptoBilis to pause all sales and shipments of Ledger devices. We recommend Ledger users who purchased from this reseller in the last 90 days to not initiate set up if you have not done so yet. If you have set up your Ledger device, consider moving assets to a new Ledger signer (with new seed). We will continue to inform customers of updates as the investigation progresses. Reach out to Ledger customer support through official channels with any questions: support.ledger.com/
1,154
1,731
1,858
7,191
4,743,215
🚨NEW EXPERIMENT 🚨 Playground is an experimental gaming platform that lets you create your own games with zero coding experience. If you can think it, you can play it. Go to labs.google/playground to learn more! Available to users 18+ in the US.
1,197
3,665
1,630
34,464
7,068,658
ivansc retweeted
The Claude Startups program is expanding to more founders. Members can get a year of Claude Team, API credits, special offers on the tools a startup runs on, and office hours with Anthropic’s Applied AI team.
746
1,227
918
18,794
5,899,060
很多人都在說AI會不會失控毁滅人類?AI毁滅人類的機率有多少? 有些AI頂尖專家也提出了各不相同的機率。 其實AI專家不知道的是這機率不是來自他們所想的AI發展本身,而是來自人類的集體意識與覺醒。 如果持續貪婪追求科技與競爭,像亞特蘭提斯文明走向滅亡就不是機率而是必然。
2
72
日本正式批准將其價值 8 兆美元的政府債券市場整個上鏈,實現 24/7 即時結算…
🇯🇵 Japan officially moves to put its $8 TRILLION government bond market on blockchain. The Financial Services Agency, Ministry of Finance, Bank of Japan, and the country's three megabanks are building a system for INSTANT 24/7 settlement of stocks and government bonds, per Nikkei. JPMorgan and BlackRock Japan are already testing rival blockchain settlement systems for the same market. This comes weeks after Japan ended its 4-year crypto freeze and moved to treat digital assets as financial products.
1
2
90
ivansc retweeted
Last night Apple briefly removed Telegram from the AppStore because a user had planted illegal porn in a public chat. Telegram was restored within hours. But I want to explain what happened — to warn other app developers and help protect online communities from similar attacks. Because Telegram quickly removes illegal content from public groups using all kinds of moderation tools, the attacker had to resort to a technical trick. He inserted AI-modified illegal content by editing an old message in an active group chat. As a result the content was effectively hidden from the group’s members, preventing them from seeing/reporting it. The attacker was a takedown extortionist: someone who demands ransom from group owners in exchange for not targeting their communities. These extortionists use automated accounts to plant illegal content in public groups and then report it directly to Apple, attempting to trigger the removal of legitimate communities whose owners refused to pay them. From a practical standpoint, illegal pornographic content in Telegram’s public groups is not a systemic problem. Our moderation is effective. The fact that attackers must resort to backdated, effectively invisible content and other technical tricks proves this. ⚠️ However, there are two important lessons here for app developers and online communities: — Extortionists have found a way to manipulate Apple into overreacting. Apple removed Telegram from the App Store before contacting us. This creates a potential systemic risk for every mobile app that hosts user-generated content. If an app used by more than a billion people can be removed from the App Store without prior warning, any app can be. — The tactics used by takedown extortionists are evolving, putting communities across social platforms at risk. Telegram has extensive experience identifying the tricks used by coordinated reporting gangs and protecting legitimate communities (even when doing so risks our own app being temporarily removed from the App Store). Other platforms may not be equally prepared. Stay vigilant! ☝️
1,461
3,691
496
33,406
4,281,539
ivansc retweeted
On several coding and knowledge work evaluations, Opus 5 is the new state-of-the-art:
467
986
1,491
10,657
5,842,473
這幾天看到很多人說 @Kimi_Moonshot 的K3有多厲害,卻沒有人注意到自己使用Kimi的資料,可能被拿去做訓練的素材,除了用Kimi的API有明確指出不會用來做訓練,其他的使用方式都沒有這方面的說明,也沒有開關可以控制保護自己的隱私。 如果是企業或注重隱私的,應該要多注意,而不是只看模型給的成果或成本。 這部分 Claude和Codex在不同方案的隱私差異說明都看得到。 希望 @Kimi_Moonshot 能在這部分做一個明確的說明,消除隱私的疑慮。
4
65