@it-industry-news@baudrate.tw
Search for this handle on your own server to follow from there.
It would be “good for the world” to slow down the pace of AI development, according to a blog post from Anthropic, which this week began the process of going public with a confidential IPO filing. “We…
新聞在Computex 2026期間，Nvidia與微軟公布Windows個人AI代理運算平臺Nvidia RTX Spark，採用Nvidia Blackwell RTX GPU、20核心Nvidia Grace CPU與最高128GB統一記憶體架構（Unified Memory），鎖定創作者、AI開發者與玩家，主打特色是大型模型與AI代理可在使用者端裝置執行。Nvidia表示，RTX Spar…
新聞AI新創Anthropic近日發表報告《When AI builds itself》，揭露該公司合併至正式程式碼庫的程式碼中，已有超過80%由Claude撰寫，顯示AI正逐步參與AI模型本身的開發流程。Anthropic並主張，全球應建立可驗證的協調機制，讓主要前沿AI開發者在必要時具備共同減速或暫停開發的選項。Anthropic表示，AI模型的開發流程正在快速改變。早期Claude的開發工作…
新聞在2025年，群創光電悄悄完成一項組織變革，對於群創光電、對於臺灣製造業的資安史上，這是一個值得記錄的里程碑。群創光電的資通安全部正式從MIS（資訊系統）部門獨立出來，直屬執行長辦公室，並設立資安長（CISO）一職，邀請學者型高階主管林宜隆擔綱。這個組織圖上的移動，看似只是幾條線的重新連接，卻牽動一整套思維方式的根本翻轉——資訊安全，不再只是IT部門底下一個防堵漏洞的執行單位，而是公司戰略決策…
Bees Can Use Tools To Solve Problems, Study Finds https://science.slashdot.org/story/26/06/04/2016228/bees-can-use-tools-to-solve-problems-study-finds?utm_source=rss1.0mainlinkanon
新聞資安漏洞情報業者VulnCheck於5月17日觀察到，殭屍網路RondoDox開始利用華碩（Asus）路由器漏洞CVE-2018-5999。這項漏洞早在2018年已有公開利用程式，但VulnCheck表示，這是該公司首度觀察到實際濫用跡象。BankInfoSecurity報導指出，這項漏洞可讓未經身分驗證的攻擊者以root權限遠端執行程式碼；VulnCheck技術長Jacob Baines也對…
新聞資安公司Adversa AI揭露SymJack新型攻擊鏈，協助開發者寫程式、修改檔案與執行命令的AI程式開發代理。研究團隊指出，攻擊者可準備一個惡意程式碼儲存庫，讓AI代理執行看似普通的檔案複製動作，但實際內容可能被寫入代理工具的設定檔，並在代理下次啟動時執行攻擊者指定的程式。研究團隊已在Claude Code、Gemini CLI與Antigravity CLI、Cursor Agent C…
新聞宏碁本週發布安全公告，預告將針對Wave 7家用路由器系列發布韌體更新，以修補2個CVSS風險值10.0的安全漏洞。Wave 7為宏碁的家用Wi-Fi 7 Mesh網狀網路路由器。宏碁經由外部安全研究人員Gergo Pap通報獲知兩項與存取控制結構及韌體密碼防護機制有關的漏洞，公司正在開發韌體更新，將在完成後提供下載更新。這兩項漏洞編號分別為CVE-2026-49200及CVE-2026-49…
Benchmarking The BORE Scheduler Performance With CachyOS Linux Earlier this week I ran benchmarks of different CachyOS Linux kernel flavors that proved interesting from the performance overhead of the…
Phoronix@masto.ai
·
We’re a month shy of the 250th anniversary of the United States’ independence, and another close ally has decided to celebrate by declaring independence from American tech, AI in particular. The Canad…
UPDATED A new extortion brand called Pink – which may be a rebrand of BlackFile – uses voice phishing and fake help-desk calls to gain initial access to organizations’ IT environments, steal their sen…
Anthropic Urges Global Pause in AI Development, Flags 'Self-Improvement' Risk https://slashdot.org/story/26/06/04/204255/anthropic-urges-global-pause-in-ai-development-flags-self-improvement-risk?utm_…
Valve says it’s ready to launch the Steam Machine this summer https://www.theverge.com/games/943657/valve-steam-machine-frame-summer-launch-verified
New IronWorm Malware Hits 36 Packages In npm Supply-Chain Attack https://it.slashdot.org/story/26/06/04/1948205/new-ironworm-malware-hits-36-packages-in-npm-supply-chain-attack?utm_source=rss1.0mainli…
Linux 7.1 + Mesa 26.1 Performance With The Radeon RX 9070 GRE, RX 9070 XT With this week's launch day review of the AMD Radeon RX 9070 GRE, Ubuntu 26.04 with its Linux 7.0 and Mesa 26.0 default driver…
Phoronix@masto.ai
·
AT&T and Verizon lose Supreme Court case over fines for selling location data FCC did not violate carriers' right to jury trial, court says in 8-1 ruling. https://arstechnica.com/tech-policy/2026/06/a…
Companies Are Using Reddit To Manipulate ChatGPT and Google AI Search https://tech.slashdot.org/story/26/06/04/1828244/companies-are-using-reddit-to-manipulate-chatgpt-and-google-ai-search?utm_source=…
These LLMs are the best at resisting Russian propaganda Estonian government benchmark shows how dozens of models combat Russia's "strategic narratives." https://arstechnica.com/ai/2026/06/these-llms-a…
Cyberdecks used to look like little laptops, but now they’re getting more personal https://www.theverge.com/tech/943445/cyberdeck-tiktok
Meta Keeps Delaying the Release of Its New AI Model to Developers https://meta.slashdot.org/story/26/06/04/181247/meta-keeps-delaying-the-release-of-its-new-ai-model-to-developers?utm_source=rss1.0mai…
