Hacker News

Article URL: https://blog.skacekamen.dev/posts/nest-js-is-not-really-good/

Comments URL: https://news.ycombinator.com/item?id=47297383

Points: 1

# Comments: 0

I'm looking for an arXiv endorser in the cs.DS (Data Structures) or cs.DB category.

I've spent the past week developing a novel hierarchical data structure called the Omni Axis Tree (OAT) — a multi-parent, multi-dimensional tree with O(1) exact retrieval via bidirectional indexing. It achieves 193× faster retrieval than DAG traversal and 64× faster than vector search at 50,000 nodes, with retrieval time staying flat as dataset size grows.

The structure has applications in AI agent memory systems, enterprise knowledge graphs, and cross-dimensional analytics.

I have a research paper ready for submission and a provisional patent filed (U.S. App. No. 63/999,482).

If you've published on arXiv in cs.DS or cs.DB and would be willing to endorse my submission, I'd be very grateful. Happy to share the paper first so you can evaluate it.

Feel free to DM me or drop a comment below.

Thank you

Comments URL: https://news.ycombinator.com/item?id=47297373

Points: 1

# Comments: 0

I built an autonomous AI operating system that runs locally on Linux.

It's not a chatbot — it decomposes complex tasks into subtasks, executes them autonomously, writes and runs code, browses the web, and learns from every interaction.

Key features: - Multi-model support: Claude, GPT-4, Gemini, Groq, DeepSeek (any OpenAI-compatible) - Project Manager: breaks down projects into subtasks and executes them in parallel - Persistent memory (vector + SQLite) - DeepThink: secondary reasoning layer that reviews critical actions before execution - Plugin system: drop a .py file into plugins/ and it's instantly available - Real-time Live Activity window showing every tool call as it happens

Built with Python + PyQt6. MIT license.

GitHub: https://github.com/FiratBulut/Moruk-OS

Comments URL: https://news.ycombinator.com/item?id=47297364

Points: 1

# Comments: 0

Article URL: https://www.thelightphone.com/shop/light-phone-iii-accessories

Comments URL: https://news.ycombinator.com/item?id=47297352

Points: 1

# Comments: 0

Article URL: https://omnifish.ee/eclipse-glassfish-this-isnt-your-fathers-glassfish/

Comments URL: https://news.ycombinator.com/item?id=47297346

Points: 1

# Comments: 0

Article URL: https://offsec.almond.consulting/bypassing-apache-fop-escaping-to-reach-ghostscript.html

Comments URL: https://news.ycombinator.com/item?id=47297326

Points: 1

# Comments: 0

Article URL: https://papervault.xyz/

Comments URL: https://news.ycombinator.com/item?id=47297325

Points: 2

# Comments: 1

I built a dashboard that tries to make country-level economic data actually understandable.

Instead of "$X billion in exports," it shows Italy shipped 96M dresses, 11K yachts, and 2.1B liters of wine.

Screenshots: https://imgur.com/a/kuxWr55

Covers 37 countries across trade (UN Comtrade), government spending (OECD/COFOG), revenue, debt, strategic resource dependencies, and 24 World Bank indicators.

Stack: Next.js + React + D3.js, FastAPI + SQLite backend.

Not live yet, sharing screenshots for feedback before launch. Curious if this is useful to anyone, or if I'm building for an audience of one.

Comments URL: https://news.ycombinator.com/item?id=47297321

Points: 1

# Comments: 0

Article URL: https://am.jpmorgan.com/us/en/asset-management/liq/insights/market-insights/eye-on-the-market/energy-paper-2026/

Comments URL: https://news.ycombinator.com/item?id=47297315

Points: 1

# Comments: 0

I built a browser-based tool that uses Vision-Language Models (VLMs) to detect objects in satellite imagery via natural language prompts. Draw a polygon on the map, type what you want to find (e.g., "swimming pools," "oil tanks," "solar panels"), and the system scans tile-by-tile, projecting bounding boxes back onto the globe as GeoJSON. The pipeline: pick zoom level + prompt → slice map into mercantile tiles → feed each tile + prompt to VLM → create bounding boxes → project to WGS84 coordinates → render on map. No login required for the demo. Works well for distinct structures zero-shot; struggles with dense/occluded objects where narrow YOLO models still win.

Comments URL: https://news.ycombinator.com/item?id=47297308

Points: 1

# Comments: 0

Article URL: https://twitter.com/theflow0/status/2030011206040256841

Comments URL: https://news.ycombinator.com/item?id=47297288

Points: 1

# Comments: 0

Article URL: https://github.com/vijaym2k6/SteerPlane

Comments URL: https://news.ycombinator.com/item?id=47297274

Points: 1

# Comments: 0

Article URL: https://twitter.com/angaisb_/status/2029635731585372598

Comments URL: https://news.ycombinator.com/item?id=47297271

Points: 2

# Comments: 0

I kept running lsof -i :PORT repeatedly to check what was alive locally. Eventually I wrote a small tool around it and it turned into this.

How it works: - One lsof / netstat call maps all listening processes - 100+ ports scanned in parallel via TCP connect, 150ms timeout each - HTTP fingerprinting reads response headers to identify frameworks (Spring Boot, Express, Django, etc.) - Filesystem detection reads package.json, pom.xml, go.mod, Cargo.toml from the process working directory to get the actual project name - Redis gets a PING/PONG health check, Postgres gets a connection handshake - Results pushed to the browser via WebSocket every few seconds

First scan: ~800ms. Subsequent scans: ~300-400ms. Memory: ~55MB RSS.

Detects 40+ services including databases, queues, monitoring tools, and AI runtimes (Ollama).

Single file UI, no build step, no bundler. Just HTML/CSS/JS served by Express.

npm install -g @haseeb-xd/havn && havn

GitHub: https://github.com/haseeb-xd/havn

Happy to answer questions about the detection approach or the performance tradeoffs.

Comments URL: https://news.ycombinator.com/item?id=47297233

Points: 1

# Comments: 0

Article URL: https://www.invoica.ai/

Comments URL: https://news.ycombinator.com/item?id=47297225

Points: 1

# Comments: 1

Article URL: https://www.macrumors.com/2026/03/07/apple-adds-three-executives-to-leadership-page/

Comments URL: https://news.ycombinator.com/item?id=47297220

Points: 2

# Comments: 2

Article URL: https://niji.webs.me/blog/2026-03-08-how-claude-code-compresses-context.html

Comments URL: https://news.ycombinator.com/item?id=47297211

Points: 3

# Comments: 1

Article URL: https://floaty.one

Comments URL: https://news.ycombinator.com/item?id=47296966

Points: 1

# Comments: 0

Article URL: https://kiss-and-solid.com/blog/keep-it-simple

Comments URL: https://news.ycombinator.com/item?id=47296963

Points: 1

# Comments: 0

An Agent's core capability comes entirely from the underlying LLM. Therefore, the future of Agents is strictly dictated by the present state of LLMs.

So, where exactly are LLMs right now?

I believe we are currently in the "cottage industry" (or subsistence) phase of AI—the very dawn of industrialization. To use a historical analogy: we just invented the first steam engines. They are bulky, stationary, and only good for pumping water out of coal mines. We aren't anywhere close to having steam locomotives yet.

Right now, there's a massive explosion of custom Agents being built. But if you look closely, they are almost entirely "self-sufficient" and siloed. Everyone is building their own Agent for their own specific use case, but it's incredibly hard to adapt or scale them for broader use. It’s like every household having its own loom, weaving its own cloth, and never using anyone else's.

Why is this happening? It comes down to the current limits of LLMs. If we put aside multimodal capabilities for a moment, text-based LLMs basically have four core pillars:

Natural Language Understanding (NLU)

Natural Language Generation (NLG)

Tool Calling

Reasoning

The first three are already highly mature and reliable. But the fourth—Reasoning—is still an absolute minefield of hallucinations.

Yet, what do Agent developers obsess over the most? Reasoning. Why? Because it looks cool on a demo. This obsession is exactly why we can't truly "industrialize" Agents yet. It’s why it is so damn hard to find a genuinely reliable, general-purpose Agent in the wild (the recent hype and reality check around Manus is a textbook example of this).

Sure, one day LLM reasoning capabilities might surpass 99% of humanity. When that day comes, we will finally see truly powerful, general-purpose Agents. But honestly, nobody knows exactly when that timeline will hit.

My takeaway: If I am building a general Agent for production today, I am strictly utilizing NLU, NLG, and Tool Calling. I am staying the hell away from relying on "Reasoning."

A recent convo with some friends about AI got me thinking. My take seemed to resonate with them, so I’m sharing it here to hear your thoughts

Comments URL: https://news.ycombinator.com/item?id=47296947

Points: 1

# Comments: 0