<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: jdeng</title><link>https://news.ycombinator.com/user?id=jdeng</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 19 Nov 2026 15:24:42 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=jdeng" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by jdeng in "Native Shell (NSH) is a native UI framework. Build once, native everywhere"]]></title><description><![CDATA[
<p>Great work! The gallery demo is surprisingly beautiful compared to other cross platform native UI framework.</p>
]]></description><pubDate>Tue, 10 Nov 2026 09:44:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=50438030</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=50438030</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50438030</guid></item><item><title><![CDATA[New comment by jdeng in "The agent harness belongs outside the sandbox"]]></title><description><![CDATA[
<p>Good comparison of the 2 approaches of agent harness. Vita AI[0] also picked the outside model. We used the E2B sandbox solution, and fortunately it supports volume, so we just create volumes for personal skills and organization workspace, and mount them to the sandbox. This elegantly solves the "file system" problem you mentioned.<p>[0]: <a href="https://www.vita-ai.net" rel="nofollow">https://www.vita-ai.net</a> (AI agent for content creation & social media management)</p>
]]></description><pubDate>Sun, 03 May 2026 15:42:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=47998078</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=47998078</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47998078</guid></item><item><title><![CDATA[New comment by jdeng in "DeepSeek v4"]]></title><description><![CDATA[
<p>Excited that the long awaited v4 is finally out. But feel sad that it's not multimodal native.</p>
]]></description><pubDate>Fri, 24 Apr 2026 03:53:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=47885300</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=47885300</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47885300</guid></item><item><title><![CDATA[AI makes everything easy to build and impossible to sell]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.vita-ai.net/blogs/ai-creator-economy-disruption">https://www.vita-ai.net/blogs/ai-creator-economy-disruption</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47425183">https://news.ycombinator.com/item?id=47425183</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 18 Mar 2026 12:52:59 +0000</pubDate><link>https://www.vita-ai.net/blogs/ai-creator-economy-disruption</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=47425183</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47425183</guid></item><item><title><![CDATA[New comment by jdeng in "Show HN: Klaus – OpenClaw on a VM, batteries included"]]></title><description><![CDATA[
<p>For openclaw to become helpful, you have to connect it to your personal email, access to your file etc. All of these requires user's manual setup right?. I do not get the point of "batteries included". Installing it is not the bottleneck right? The official docs has detail procedures for all deployment options.</p>
]]></description><pubDate>Thu, 12 Mar 2026 13:56:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=47350591</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=47350591</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47350591</guid></item><item><title><![CDATA[The Three Eras of Applications: From Desktop to Cloud to AI-Native]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.vita-ai.net/blogs/three-eras-of-applications">https://www.vita-ai.net/blogs/three-eras-of-applications</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47311060">https://news.ycombinator.com/item?id=47311060</a></p>
<p>Points: 1</p>
<p># Comments: 0</p>
]]></description><pubDate>Mon, 09 Mar 2026 16:17:09 +0000</pubDate><link>https://www.vita-ai.net/blogs/three-eras-of-applications</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=47311060</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47311060</guid></item><item><title><![CDATA[New comment by jdeng in "Show HN: Mog, a programming language for AI agents"]]></title><description><![CDATA[
<p>Good try. I had the idea before, but later realized all the agent ecosystem has already highly coupled with bash scripts including skills. Though bash is not a good programming language, it's indeed flexible for CLI apps. I also realized that what agent and skills really need is just a glue language, not a real programming language, so bash is not so bad in the case.</p>
]]></description><pubDate>Sat, 07 Mar 2026 03:47:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=47284292</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=47284292</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47284292</guid></item><item><title><![CDATA[New comment by jdeng in "Ask HN: What are you working on? (February 2026)"]]></title><description><![CDATA[
<p>Working on a Vibe Testing Agent (<a href="https://www.vita-ai.net/" rel="nofollow">https://www.vita-ai.net/</a>), a perfect companion of vibe coding. Now the loop is vibe coding <-> vibe testing.</p>
]]></description><pubDate>Tue, 10 Feb 2026 02:41:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=46954668</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=46954668</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46954668</guid></item><item><title><![CDATA[New comment by jdeng in "Kimi Released Kimi K2.5, Open-Source Visual SOTA-Agentic Model"]]></title><description><![CDATA[
<p>Glad to to see open source models are catching up and treat vision as first-class citizen (a.k.a native multimodal agentic model). GLM and Qwen models takes different approach, by having a base model and a vision variant (glm-4.6 vs glm-4.6v).<p>I guess after Kimi K2.5, other vendors are going to the same route?<p>Can't wait to see how this model performs on computer automation use cases like
VITA AI Coworker.<p><a href="https://www.vita-ai.net/" rel="nofollow">https://www.vita-ai.net/</a></p>
]]></description><pubDate>Tue, 27 Jan 2026 10:52:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=46778313</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=46778313</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46778313</guid></item><item><title><![CDATA[Show HN: Vita AI Coworker – Autonomous agents for testing and desktop automation]]></title><description><![CDATA[
<p>I've been following the Claude Cowork discussion this week with interest. The "AI coworker" framing resonated—we've been building Vita AI around the same idea: agents that work independently, not assistants you supervise.<p>We're focused on two main use cases: testing web apps and desktop automation. The key difference: our agents run in persistent cloud sandboxes instead of on your local machine.<p>Why this matters:<p><i>Isolation</i> – Your local files and network never touch the agent. No DNS exfiltration, no filesystem access to your machine. The sandbox runs in the cloud, persistent for each task but completely isolated from your computer.<p><i>True autonomy</i> – Tell it "test the checkout flow" or "automate this browser workflow," close your laptop, come back hours later to a full report with screenshots and logs. You delegate the work and walk away. No babysitting required.<p><i>Scale</i> – Spin up 10 parallel agents without melting your CPU. Each agent gets its own environment in the cloud.<p>Trade-off: you can't use it offline, and there's network latency. But for testing and automation, async cloud execution is actually what you want.<p>Try it at <a href="https://vita-ai.net" rel="nofollow">https://vita-ai.net</a> (free trial available). Built with Next.js 15, Drizzle ORM, and Vercel's AI SDK.<p>Different use case than Cowork (specialized for testing/automation vs. general tasks), but the same philosophical shift: moving from assistants to autonomous agents.<p>Happy to answer questions about the architecture or how we think about security.</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=46659144">https://news.ycombinator.com/item?id=46659144</a></p>
<p>Points: 1</p>
<p># Comments: 0</p>
]]></description><pubDate>Sat, 17 Jan 2026 16:09:50 +0000</pubDate><link>https://www.vita-ai.net</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=46659144</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46659144</guid></item><item><title><![CDATA[New comment by jdeng in "Cowork: Claude Code for the rest of your work"]]></title><description><![CDATA[
<p>Exciting to see Anthropic validate the "AI coworker" direction. We're building VITA AI (<a href="https://vita-ai.net" rel="nofollow">https://vita-ai.net</a>) with similar philosophy but for enterprise QA testing.<p>One key architectural difference: Cowork runs sandboxed VMs on your local macOS machine, but we run sandboxes entirely in the cloud. This means:<p>- <i>True isolation</i> - agents never touch your local files or network, addressing the security concerns raised in this thread<p>- <i>Actual autonomy</i> - close your laptop, agent keeps working. Like delegating to a real coworker, not pairing with an assistant<p>- <i>Scale</i> - spin up 10 test agents without melting your CPU<p>The trade-off is latency and offline capability, but for testing workflows (our focus), asynchronous cloud execution is actually the desired model. You assign "test the checkout flow," go to lunch, come back to a full test report + artifacts.<p>Different use cases, different architectures. But the broader trend feels right - moving from conversational assistants to autonomous agents that operate independently.</p>
]]></description><pubDate>Fri, 16 Jan 2026 16:11:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=46648017</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=46648017</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46648017</guid></item><item><title><![CDATA[New comment by jdeng in "Ask HN: What Are You Working On? (Nov 2025)"]]></title><description><![CDATA[
<p>Sure. Join the waitlist below and I will note down that you need private demo. I will contact you once we got bandwidth on that (private demo).<p><a href="https://www.vita-ai.net/waitlist" rel="nofollow">https://www.vita-ai.net/waitlist</a></p>
]]></description><pubDate>Tue, 11 Nov 2025 01:44:50 +0000</pubDate><link>https://news.ycombinator.com/item?id=45883178</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=45883178</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45883178</guid></item><item><title><![CDATA[New comment by jdeng in "Ask HN: What Are You Working On? (Nov 2025)"]]></title><description><![CDATA[
<p>Working on an AI coworker (QA engineer) that can autonomously tests web app like your human QA colleagues. It has its own virtual desktop and runs as background agent.<p><a href="https://www.vita-ai.net/" rel="nofollow">https://www.vita-ai.net/</a></p>
]]></description><pubDate>Mon, 10 Nov 2025 13:57:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=45876057</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=45876057</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45876057</guid></item><item><title><![CDATA[Show HN: MCPBar – Open-Source Package Manager with Open Registry for MCP Servers]]></title><description><![CDATA[
<p>I've built MCPBar (<a href="https://www.mcp.bar" rel="nofollow">https://www.mcp.bar</a>), an open-source CLI tool that brings npm-like package management to the Model Context Protocol (MCP) ecosystem, solving the fragmentation problem for AI tools.<p>What is MCPBar?<p>- A package manager for discovering and installing MCP servers (which connect AI models to tools/APIs)<p>- Uses a standardized mcp.json manifest format (similar to package.json) (link: <a href="https://github.com/in-fun/mcpbar/blob/main/doc/proposal.md">https://github.com/in-fun/mcpbar/blob/main/doc/proposal.md</a>)<p>- Works across multiple MCP clients like Claude, Cursor, and Windsurf<p>- Includes an open registry with standardized metadata for MCP servers<p>Why I built it:<p>The MCP ecosystem is growing rapidly but lacks standardization. Each AI client implements MCP differently, making discovery and installation frustratingly manual. MCPBar creates a decentralized-yet-standardized approach to package management for AI tools.<p>Technical highlights:<p>- Decentralized registry pattern inspired by npm and ESM imports<p>- Manifest files contain standardized metadata and installation instructions<p>- Cross-platform support (works on macOS/Windows/Linux)<p>- Simple CLI interface: mcpbar install github/github-mcp-server<p>Try it:<p>npm install -g mcpbar<p>mcpbar search github    # Find GitHub-related MCP servers<p>mcpbar install github/github-mcp-server   # Install a GitHub MCP server<p>GitHub: <a href="https://github.com/in-fun/mcpbar">https://github.com/in-fun/mcpbar</a><p>Project site: <a href="https://www.mcp.bar" rel="nofollow">https://www.mcp.bar</a><p>I'd love feedback on the manifest format, the CLI interface, and whether this approach makes sense for the evolving MCP ecosystem. Are there other features you'd want this to support?</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=43906473">https://news.ycombinator.com/item?id=43906473</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Tue, 06 May 2025 15:49:27 +0000</pubDate><link>https://www.mcp.bar</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=43906473</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43906473</guid></item><item><title><![CDATA[New comment by jdeng in "Show HN: Rabrain – AI Note Taking with Voice Input and Notion-Like Editor"]]></title><description><![CDATA[
<p>Thanks! Voice input and AI assisted notes really save me. Whenever I have some spark, I just talk to AI and ask it to transform my thoughts into notes.</p>
]]></description><pubDate>Tue, 08 Apr 2025 14:49:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=43622437</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=43622437</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43622437</guid></item><item><title><![CDATA[Show HN: Rabrain – AI Note Taking with Voice Input and Notion-Like Editor]]></title><description><![CDATA[
<p>Hello HN community!<p>I'm thrilled to introduce Rabrain, a tool that redefines how you manage and recall information online. Rabrain combines the best of bookmark management and AI note-taking, featuring a Notion-like editor with powerful voice input capabilities.<p>Key Features:<p>One-Click Save: Easily save articles, PDFs, and web pages with our Chrome extension.<p>AI Summarization: Get concise summaries of long articles in seconds.<p>Voice Input: Capture your thoughts on the go with voice-to-text capabilities.<p>AI-Assisted Note Taking: Use AI to draft, refine, and organize your notes effortlessly.<p>Rabrain is perfect for anyone looking to enhance their productivity and knowledge management. Whether you're a student, researcher, or professional, Rabrain helps you keep track of valuable information and recall it exactly when you need it.<p>I built Rabrain to solve my own frustrations with managing digital information, and I'm eager to hear your thoughts and feedback. Check it out and let me know what you think!<p>Link: <a href="https://www.rabrain.com" rel="nofollow">https://www.rabrain.com</a><p>Looking forward to your comments and suggestions!</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=43621953">https://news.ycombinator.com/item?id=43621953</a></p>
<p>Points: 3</p>
<p># Comments: 2</p>
]]></description><pubDate>Tue, 08 Apr 2025 14:06:59 +0000</pubDate><link>https://www.rabrain.com</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=43621953</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43621953</guid></item><item><title><![CDATA[Show HN: SnapDrop – An AirDrop alternative that works on any moderm web browsers]]></title><description><![CDATA[
<p>Recently snapdrop.net and sharedrop.io had been acquired (hijacked) by LimeWire. The widely used local file sharing feature had become "upload files". This is really annoying, since I have trusted the sites for years. Before I noticed the hijack, I even accidentally uploaded several secret files to it.<p>To avoid others falling to the same trap as I was, I deployed this SnapDrop instance that is clean and free.<p># SnapDrop Features<p>File sharing on your local network that works on all platforms.<p>A cross-platform AirDrop-like solution that works<p>Send images, documents or text via peer-to-peer connection to devices on the same local network.<p>Internet transfers<p>Join temporary public rooms to transfer files easily over the Internet.<p>Web app<p>Works on all devices with a modern web browser.<p># SnapDrop Use Cases<p>- Send a file from your phone to your laptop<p>- Share photos in original quality with friends using Android and iOS<p>- Share private files peer-to-peer between Linux systems</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=43286812">https://news.ycombinator.com/item?id=43286812</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Fri, 07 Mar 2025 01:48:08 +0000</pubDate><link>https://www.snap-drop.net</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=43286812</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43286812</guid></item><item><title><![CDATA[What do you think of a new language as better OCaml?]]></title><description><![CDATA[
<p>Article URL: <a href="https://github.com/in-fun/hi">https://github.com/in-fun/hi</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=41952496">https://news.ycombinator.com/item?id=41952496</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Sat, 26 Oct 2024 04:12:56 +0000</pubDate><link>https://github.com/in-fun/hi</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=41952496</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41952496</guid></item><item><title><![CDATA[New comment by jdeng in "Black Mirror: is deep thinking on the decline?"]]></title><description><![CDATA[
<p>>A bow is more powerful than a stick but also develops more motor skills, a bike develops the users body more than a car, and so on.<p>Totally agree. I like bike much more than car, however the trend of bike superseded by car cannot be reversed, given car's convenience and power.<p>IMO if we cannot reverse it, then embrace it. It applies to other tech advancement as well.</p>
]]></description><pubDate>Tue, 27 Aug 2024 01:21:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=41363803</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=41363803</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41363803</guid></item><item><title><![CDATA[New comment by jdeng in "Black Mirror: is deep thinking on the decline?"]]></title><description><![CDATA[
<p>The decline of deep thinking in the digital age is a pressing concern. While technology presents challenges to our cognitive faculties, it also offers unprecedented opportunities. By leveraging AI as a cognitive partner, we can counteract the negative effects of constant connectivity and revive our capacity for deep, meaningful thought.</p>
]]></description><pubDate>Tue, 27 Aug 2024 00:39:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=41363595</link><dc:creator>jdeng</dc:creator><comments>https://news.ycombinator.com/item?id=41363595</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41363595</guid></item></channel></rss>