惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

V
Vulnerabilities – Threatpost
V
V2EX
GbyAI
GbyAI
Recent Announcements
Recent Announcements
Microsoft Security Blog
Microsoft Security Blog
阮一峰的网络日志
阮一峰的网络日志
Hugging Face - Blog
Hugging Face - Blog
T
Tailwind CSS Blog
Y
Y Combinator Blog
C
Check Point Blog
爱范儿
爱范儿
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
美团技术团队
雷峰网
雷峰网
IT之家
IT之家
WordPress大学
WordPress大学
V
Visual Studio Blog
Microsoft Azure Blog
Microsoft Azure Blog
MyScale Blog
MyScale Blog
N
News and Events Feed by Topic
罗磊的独立博客
S
SegmentFault 最新的问题
S
Security Affairs
aimingoo的专栏
aimingoo的专栏
F
Fortinet All Blogs
K
KPMG report finds enterprise disconnect between AI and its ROI | CIO
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
H
Hacker News: Front Page
Google DeepMind News
Google DeepMind News
B
Blog
O
OpenAI News
C
Cisco Blogs
Simon Willison's Weblog
Simon Willison's Weblog
The Last Watchdog
The Last Watchdog
Hacker News: Ask HN
Hacker News: Ask HN
博客园_首页
人人都是产品经理
人人都是产品经理
C
Cybersecurity and Infrastructure Security Agency CISA
Recent Commits to openclaw:main
Recent Commits to openclaw:main
Help Net Security
Help Net Security
月光博客
月光博客
J
Java Code Geeks
L
LangChain Blog
博客园 - 司徒正美
Stack Overflow Blog
Stack Overflow Blog
Security Archives - TechRepublic
Security Archives - TechRepublic
Apple Machine Learning Research
Apple Machine Learning Research
T
The Exploit Database - CXSecurity.com
N
News and Events Feed by Topic
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More

Hacker News - Newest: "AI"

AI can't read an investor deck AI as an attorney? Student uses ChatGPT, Gemini to sue UW over alleged racial discrimination Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval GitHub - MeepCastana/KubeezCut: Free Web based video editor GitHub - GenAI-Gurus/awesome-eu-ai-act: Curated tools, official sources, OSS, templates, and guides for EU AI Act compliance. Can AI judge journalism? A Thiel-backed startup says yes, even if it risks chilling whistleblowers Coming soon: 10 Things That Matter in AI Right Now DARPA built an AI to fact-check enemy weapons claims What explains heterogeneity in AI adoption? When AI Meets Muscle: Context-Aware Electrical Stimulation Promises a New Way to Guide Human Movements - Department of Computer Science AI Changed How We Build. It Did Not Change What Matters. Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Meta spins up AI version of Mark Zuckerberg to engage with employees Code Mode: Let Your AI Write Programs, Not Just Call Tools | TanStack Blog GitHub - Delavalom/graft: Go framework for building AI agents. Type-safe tools, multi-provider (OpenAI, Anthropic, Gemini, Bedrock), zero vendor SDKs. India's TCS tops estimates, says new AI models did not dent services demand Gen Z's fading AI hype Strong feeling: we are in a folded AI reality GitHub - machinarii/total-recall-catalog: A reference catalog of latest knowledge retrieval, memory & RAG systems GitHub - mensfeld/code-on-incus: Give each AI agent its own isolated machine with root, Docker, and systemd. Active defense detects and stops threats automatically.. Quantization, LoRA, and the 8% Problem: Benchmarking Local LLMs for Production AI Iran war: We spoke to the man making Lego-style AI videos that experts say are powerful propaganda Powell, Bessent discussed Anthropic's Mythos AI cyber threat with major U.S. banks GitHub - immartian/bellamem: Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact. recursive-mode: The Repo-Native Operating System for AI Engineering After the attack on Sam Altman's home, will AI CEO's go on the offensive? The biggest advance in AI since the LLM Opus 4.6 vs GPT 5.4 One Prompt Unity World Generation Test “AI polls” are fake polls Client Challenge Can AI be a 'child of God'? Inside Anthropic's meeting with Christian leaders How to Switch AI Chatbots and Why You Might Want To GitHub - MattMessinger1/agentic_refund_guardrail: Safe refund policy layer for AI agents — Python + TypeScript. Same behavior, shared tests. Adam/papers/emergent_values_whitepaper.md at master · strangeadvancedmarketing/Adam Ask HN: How do you stop playing 20 questions with your AI coding tools How far can automation and AI support psychotherapy? - @theU GitHub - stagas/rtdiff: realtime git diff gui and AI-assisted commits A Mac Studio for Local AI — 6 Months Later A History of the Early Years of AI at the University of Edinburgh Why AI Coding Tools Still Feel Stuck on Localhost MSN AI Datacenters Are Becoming Strategic Targets twitter.com Penn Researchers Use AI to Surface Unreported GLP-1 Side Effects in Reddit Posts Show HN: MoodSense AI (ML and FastAPI and Gradio, Deployed on Hugging Face) Moodsense Ai - a Hugging Face Space by aman179102 AI models are terrible at betting on soccer—especially xAI Grok GitHub - xialeistudio/echoic GitHub - HimashaHerath/github-dev-wrapped: AI-powered weekly GitHub activity reports deployed to GitHub Pages GitHub - alejandrobalderas/claude-code-from-source: Architecture, patterns & internals of Anthropic's AI coding agent — reverse-engineered from source maps AI and Tech brief: Ireland ascendant GitHub - Titovilal/context0: Context0 - Never Surrender Training for a Marathon with an AI Coach: What Worked and What Didn't Cyber Pulse: Agentic Intel - Apps on Google Play I Built an AI PR Reviewer That Catches Bugs by Not Looking for Bugs Gen Z workers are so fearful AI will take their job they’re intentionally sabotaging their company’s AI rollout | Fortune How AI Is Reimagining the Game of Golf–For Both Players and Courses GitHub - nattergabriel/reseed: A CLI tool for managing and distributing agent skills across projects Is SVG the final frontier? My AI workflow evolved from prompts to a near-autonomous workflow MLSharp Help - 3DGS Viewer & Generator I put my cognitive field based AI's runtime on GitHub Is Numble the first AI-proof game? A3: Kubernetes for autonomous AI agent fleets | Emergent Principles Deepali Vyas ("The Elite Recruiter") GitHub - msmarkgu/RelayFreeLLM: A restful API designed to route user prompts to various AI model providers. Unionized ProPublica staff are on strike over AI, layoffs, and wages Unleashing the Advantage of Quantum AI We're heading for an AI-fueled 'dementia crisis,' brain scientist warns The AI-Assisted Breach of Mexico's Government Infrastructure [pdf] GitHub - stef41/lmscan: 🔍 Detect AI-generated text and fingerprint which LLM wrote it. Open-source GPTZero alternative. Zero dependencies, works offline. MSN GitHub - visionscaper/collabmem: Enabling long-term collaboration with Agentic AI - building up episodic and world model memory over time with in-context awareness We gave an AI a 3 year retail lease in SF and asked it to make a profit | Andon Labs AI Code is Hollowing Out Open Source, and Maintainers are Looking the Other Way What leaked "SteamGPT" files could mean for the PC gaming platform's use of AI AI is the boss at this retail store. What could go wrong? GitHub - Wuzu11517/agentic-proxy: Local proxy meant to help reduce With Drones, Geophysics and ArtificiaI Intelligence, Researchers Prepare to Do Battle Against Land Mines A Single Operator, Two AI Platforms, Nine Government Agencies: The Full Technical Report 在 Steam 上购买 FriedrichAI: Offline AI 立省 10% GitHub - inevolin/resume-cli: Hit Claude usage limits? Resume any AI coding session elsewhere. Switch tools at zero friction. GitHub - atripati/ark: AI Runtime Kernel — a context operating system for AI agents. Eliminates tool bloat, loads only what’s needed, and gives LLMs their reasoning space back. How to Build a Secure AI PR Reviewer with Claude, GitHub Actions, and JavaScript This Startup Wants You to Pay Up to Talk With AI Versions of Human Experts Intel Arc Pro B70 Brings 32GB VRAM to Local AI for $949 WordPress 7.0: The Good, the AI, and the Still Missing AI on the couch: Anthropic gives Claude 20 hours of psychiatry IatroBench: Pre-Registered Evidence of Iatrogenic Harm from AI Safety Measures AI Agents Know About Supabase. They Don't Always Use It Right. The history and future of AI at Google, with Sundar Pichai Inside an AI‑enabled device code phishing campaign How Meta Used AI to Map Tribal Knowledge in Large-Scale Data Pipelines AI for Systems: Using LLMs to Optimize Database Query Execution Forecasting the Economic Effects of AI Introducing Tinker: Play with AI, bring your ideas to life AI sheds light on an ancient gaming mystery People really hate AI but not as much as Iran—or Democrats | Fortune What is an AI Product Engineer? Phoebe Gates wants her $185 million AI startup to succeed with 'no ties to my privilege or my last name': 'I have a chip on my shoulder' | Fortune
Ubuntu 26.04 is the OS for the AI agentic era, says Canonical's Mark Shuttleworth - here's why
Written by · 2026-06-02 · via Hacker News - Newest: "AI"
Canonical Ubuntu Canonical Ubuntu Mark Shuttleworth
ZDNET screenshot / Canonical Ubuntu / YouTube

Follow ZDNET: Add us as a preferred source on Google.


ZDNET's key takeaways

  • Ubuntu 26.04 is designed from the ground up for AI developers.
  • The new Ubuntu Linux comes with AI-specific dev environments.
  • This Linux also comes with Rust-based memory safety built in.

In London, Canonical founder and CEO Mark Shuttleworth argued that Ubuntu 26.04, Ubuntu Linux, is the operating system for the "AI agentic era." Well, that's easy to claim, but what does Canonical have that can back up that claim?

Shuttleworth: from curl-to-bash to confined snaps

In his keynote for Ubuntu Summit 26.04, Shuttleworth framed open source as the "raw material" of the next wave of technological disruption. Specifically, he stated that the pace of AI-driven software innovation has outstripped traditional packaging and release processes. For AI, Linux users must move beyond Advanced Packaging Tool (APT) and Red Hat Package Manager (RPM) to signed, auto-updated, policy-driven snaps. Of course, snaps have long been Canonical's answer to delivering upstream software, but now AI requires updates at internet speed without sacrificing auditability or control, and that means snaps.

Also: Ubuntu 26.04 vs. Fedora 44: After years of testing both Linux distros, here's my verdict

Shuttleworth cited fresh telemetry from developer Alan Pope's Snap Store dashboard that shows dozens of snap updates landing in a single morning, across architectures from x86 and Arm to RISC-V and Power, all coming from the same tested bits. He positioned snaps, with confinement, progressive rollouts, channels, and enterprise gating, as the "single best, safest way to deliver bits to any Linux distro on the planet."

While Shuttleworth defended snaps in general, Ubuntu VP of Engineering Jon Seager drilled into new user-facing behavior: fine-grained permission prompts for snapped apps, similar to those in Android and iOS. For example, when a confined app first tries to access the camera, the desktop can now surface a prompt asking the user to grant or deny access, thanks to new plumbing from the kernel and AppArmor up through snapd and GNOME's display manager.

Sandbox everything: from snaps to LXD, Multipass, and microVMs

The other reason Ubuntu is the operating system you want for AI, according to Shuttleworth, is security. With this Ubuntu release, everything can run in a layered toolbox. Everything? Everything. It's not just apps, but AI agents and third-party software development kits (SDKs) as well. On Ubuntu today, that approach spans snap confinement, Docker/OCI containers, LXD system containers, traditional virtual machines (VMs) via Multipass, and a new generation of microVMs that blur the line between containers and virtualization.

Also: Ubuntu 26.04 surprised me - this upcoming release is seriously secure

This mix is essential, Shuttleworth claims, for "agentic engineering," where organizations may want to run thousands of agents, each believing it has a full Linux system while actually being tightly constrained for density and safety. LXD-based system containers provide the illusion of full machines for agents, while microVMs, delivered via an "Open Shell" snap that spins up hardened, per-agent environments for tools like Claude or Copilot, add hardware-enforced isolation when a kernel boundary is not enough.

Workshop: a new way to onboard developers and agents

Also, one concrete new piece is Workshop. This is a tool Canonical built on LXD to create "agentic workspaces." Worskop is meant to solve a long-standing pain point: combining sensitive developer credentials with untrusted or semi-trusted code.

Developers or teams can commit a Workshop definition to a repo. Thus, onboarding a new human or agent becomes "git clone, workshop launch." With these, the company claims, you can launch sandboxed development environments and agentic workflows that are composable and repeatable with a single command, while keeping your host system isolated.

Also: Ubuntu Core 26 offers an immutable Linux you can trust through 2041

Workshop works by booting a system container and then selectively binding in high-value secrets and resources, such as SSH keys for signed commits, access to specific datasets, and routes to remote Git servers, without dumping a developer's entire laptop environment into the sandbox. Canonical is already working with ISVs to ship signed SDKs into a dedicated Workshop store so that closed-source SDKs and agents can run alongside Ubuntu and Debian packages in a controlled environment.

AI, Ubuntu, and the 'implicit features' bet

Seager picked up where Shuttleworth left off, arguing that Canonical has no choice but to be "in the thick of" AI and agents if open source is to have any say in how these systems evolve. Rather than racing to bolt an LLM gimmick into the shell, Seager outlined a two-track strategy: implicit AI features that quietly improve existing capabilities, and explicit AI features that Canonical will roll out more cautiously.

On the implicit side, he pointed to accessibility and media as near-term opportunities: local speech-to-text, better camera autofocus, and microphone enhancement powered by small on-device models that can run even on CPU-only laptops. On the explicit side, he previewed a goal for Ubuntu 26.10: a desktop where "you can press a button and talk into any field you could previously type in," backed by models like Whisper and plumbed into every text entry surface on the system.

Also: This is my favorite Linux distro of all time - and I've tried them all

Seager was explicit that AI-driven accessibility is a core design target, not an afterthought. He said today's Linux screen readers "bluntly suck," and argued that feeding a framebuffer or camera capture into an LLM could radically improve both the description of on-screen content and the presentation of possible actions to visually impaired users.

Beyond accessibility, Seager teased "new ways of interacting with your machine" that lean on Ubuntu's existing confinement story: In an agentic desktop, each tool an agent can call would be packaged as its own confined snap, giving fine-grained control over what the agent can do on the user's behalf. He promised something concrete to "play with in the next six months," describing it as a way for non-experts to obtain "20 years of Linux desktop hacker" capability via agents, without needing the hacker background themselves.

Additionally, on the AI and HPC front, Seager stressed Canonical's work with Nvidia and AMD to make GPU enablement boring ... in a good way. Ubuntu users can now "just apt install CUDA and apt install ROCm," with Canonical and the vendors collaborating to ensure that the drivers and stacks are properly integrated and tested on 26.04.

Seager added that his own AMD GPU "has never sung as nicely as it does on 26.04" and that, for the first time, he "didn't have to endure any pain" to make it happen. Combined with Ubuntu's work on architecture variants, shipping entire archives compiled for specific instruction-set levels, such as amd64v3, Canonical wants to ensure that the expensive acceleration hardware enterprises are buying is fully supported by Ubuntu Linux and its bundled tools.

Keeping Ubuntu accessible in a token-metered world

Both Shuttleworth and Seager ended by promising to keep Ubuntu's historic promise alive, shipping "precisely the same bits" to hedge-fund quants and kids in the suburbs of Kolkata, in a world where AI usage is metered in expensive tokens. Shuttleworth warned that tying productivity and even basic understanding of code to proprietary, cloud-hosted models risks locking out the "poorest members of our digital society" unless open-weight models and open tooling remain a primary focus.

Also: 6 reasons why I've stuck with Ubuntu-based Linux distros for the last 20 years

Seager, for his part, rejected both "moral" disengagement from AI and vanity metrics like "who can spend the most tokens." He argued that open-source-savvy players like Canonical have to stay engaged, help the community work through a messy period of AI-generated "slop contributions," and guide the eventual convergence on a new generation of high-quality open-source components, now with agents and AI part of the toolkit.

Beyond AI: Rust, security, and crypto

Seager also highlighted how Ubuntu 26.04 incorporates memory safety in the base system. He highlighted three pillars: Rust-based rewrites of critical utilities, a new Rust-based cryptographic foundation called Universal Public Key Infrastructure (UPKI), and a unified, Rust-based time-sync stack.

On 26.04 Long Term Support (LTS), coreutils such as mv, cp, rm, and ls are now backed by the Rust-based uutils project, following two Canonical-funded security audits. Sudo has been replaced by sudo-rs, a Rust implementation that drops long-accumulated "ill-informed" features and tightens memory safety at the privilege boundary on every Ubuntu machine. Next, Canonical plans to swap bzip2 for a Rust implementation that Seager says is "up to 50% more efficient," with Zlib and Zstandard targeted by 28.04 -- changes he argued could translate into significant global energy savings given how widely these codecs are used.

Also: Rust will save Linux from AI, says Greg Kroah-Hartman

The point of UPKI is to bring browser-grade PKI hygiene to the Linux command line. Today, Seager noted, curl happily ignores certificate revocation lists, and command-line TLS often breaks on misconfigured certificate chains that browsers silently tolerate. UPKI will centralize revocation, intermediate preloading, and eventually post-quantum algorithms such as Merkle-tree-based schemes, with glue code being written for OpenSSL, GnuTLS, curl, and others so that tools across the stack can consume the same modern PKI data.

On the other side, a new NTP-rs utility will deliver NTP, NTS, and PTP "in a single binary, single configuration," aiming to radically simplify precision time configuration on Linux.

Speeding up Ubuntu releases for the AI age

Seager described 26.04 as the first LTS delivered under a new engineering "manifesto" that included a monthly shipping discipline enforced by an all-new release pipeline built with Go and Temporal. The team, he said, hit every monthly target, which in turn made the LTS release smoother.

Also: The third major Linux kernel flaw in two weeks has been found - thanks to AI

He added that Canonical has also been quietly rebuilding its community and communications muscle. Seager claimed that Canonical has added more core developers in the last six months than in the previous three years and has deliberately increased its blogging, Mastodon posts, podcasts, and community appearances. The result, Seager joked, is that "for anybody who doesn't like Ubuntu, it's a bit of a rough time ... you literally can't get away from us on the internet." 

It also means Ubuntu can keep up with AI's incredible pace.

Featured