Hacker News

I've been building Courtyard, a macOS desktop app designed to make local LLM workflows on Apple Silicon less tedious.

The motivation: I was tired of juggling multiple Python CLI scripts, JSONL formatting, and environment issues just to run a simple LoRA fine-tune on my Mac.

Courtyard is essentially a UI wrapper around mlx-lm combined with data preparation tools. It handles:

Dataset formatting and cleaning (privacy filtering, deduplication). Local LoRA fine-tuning via MLX on Apple Silicon. An integrated chat UI for A/B testing the base model vs. the fine-tuned adapter. Exporting to GGUF or directly to an Ollama runtime. The stack is Tauri 2.x + React + Rust + Python (mlx-lm). It's fully open-source (AGPL).

Repo: https://github.com/Mcourtyard/m-courtyard

I'd love to hear your thoughts on the architecture, MLX implementation, or any edge cases you run into. Happy to answer technical questions.

Comments URL: https://news.ycombinator.com/item?id=47146570

Points: 1

# Comments: 0

Article URL: https://www.cnbc.com/2026/02/24/anthropic-ai-hegseth-spying-defense.html

Comments URL: https://news.ycombinator.com/item?id=47146529

Points: 3

# Comments: 0

Article URL: https://techcommunity.microsoft.com/blog/adforpostgresql/microsoft-at-pgconf-india-2026/4496182

Comments URL: https://news.ycombinator.com/item?id=47146517

Points: 1

# Comments: 0

Article URL: https://www.canva.com/newsroom/news/cavalry/

Comments URL: https://news.ycombinator.com/item?id=47146510

Points: 1

# Comments: 0

Article URL: https://blog.janestreet.com/can-you-reverse-engineer-our-neural-network/

Comments URL: https://news.ycombinator.com/item?id=47146487

Points: 1

# Comments: 0

Article URL: https://www.nytimes.com/2026/02/24/technology/intel-sambanova-ceo-lip-bu-tan.html

Comments URL: https://news.ycombinator.com/item?id=47146467

Points: 2

# Comments: 0

Article URL: https://github.com/etazza2025/laravel-mbc

Comments URL: https://news.ycombinator.com/item?id=47146457

Points: 1

# Comments: 1

Article URL: https://discord.com/blog/learning-from-structure-discords-entity-relationship-embeddings

Comments URL: https://news.ycombinator.com/item?id=47146432

Points: 1

# Comments: 0

Article URL: https://www.nytimes.com/2026/02/12/science/elephant-whiskers.html

Comments URL: https://news.ycombinator.com/item?id=47146427

Points: 1

# Comments: 1

Article URL: https://rfcexplorer.net/

Comments URL: https://news.ycombinator.com/item?id=47146426

Points: 1

# Comments: 0

Article URL: https://doppelgangerdev.com/docs/software-isn-t-a-competition-it-s-a-map-of-our-frustrations

Comments URL: https://news.ycombinator.com/item?id=47146404

Points: 1

# Comments: 1

Article URL: https://vvsearch.com/workspace/news-terminal

Comments URL: https://news.ycombinator.com/item?id=47146380

Points: 1

# Comments: 0

Article URL: https://theconversation.com/will-artificial-snow-save-the-ski-industry-in-the-long-run-or-curse-it-275270

Comments URL: https://news.ycombinator.com/item?id=47146378

Points: 1

# Comments: 0

Article URL: https://ydinkin.substack.com/p/200-kernel-bugs-in-30-days

Comments URL: https://news.ycombinator.com/item?id=47146373

Points: 1

# Comments: 0

Article URL: https://hyperhell.itch.io/hyperhell

Comments URL: https://news.ycombinator.com/item?id=47146371

Points: 1

# Comments: 0

Article URL: https://www.macrumors.com/2026/02/24/tim-cook-warned-by-cia-china-taiwan-2027/

Comments URL: https://news.ycombinator.com/item?id=47146361

Points: 4

# Comments: 1

I'm not a developer — I'm the person who came up with this idea and built an interactive prototype with Claude's help. I'm posting here because this is exactly the kind of project that needs real engineers to take it somewhere.

The idea: a declare-ai.json file that any creator, platform, or AI tool can publish alongside content — declaring what percentage was written, coded, illustrated, or researched by AI versus humans, which tools were used, and who the human contributors are. A lightweight embeddable widget displays it as a collapsible pie chart. A browser extension auto-detects it on any page. A community forum handles disputes.

Think Creative Commons, but for AI provenance. Or a nutrition label, but for intelligence.

The demo is a full interactive developer briefing — architecture diagram, JSON schema, forum mockup with a real dispute thread example, tech stack recommendation, and phased roadmap. The widget on the page declares itself.

I've gifted it as MIT. I'm looking for developers who want to own this with me.

GitHub: https://github.com/Declare-AI/declare-ai Live demo: https://declare-ai.github.io/declare-ai/declare-ai-devteam.h...

Genuinely open to all feedback — including "this already exists and here's why it won't work."

Comments URL: https://news.ycombinator.com/item?id=47146360

Points: 1

# Comments: 2

Article URL: https://www.macrumors.com/2026/02/24/apple-european-iphone-sales-record-2025/

Comments URL: https://news.ycombinator.com/item?id=47146359

Points: 1

# Comments: 0

Hi HN, I'm Donald.

I’m a solo developer building NeoShift BI (and its backend companion, NeoShift ETL).

Enterprise BI tools like PowerBI or Looker are incredibly powerful, but they usually require dedicated data engineering teams to manage. On the other end of the spectrum, dumping a CSV into ChatGPT is great for a quick question but terrible for building reproducible, interactive, and shareable dashboards.

I wanted something in the middle: a lightweight BI tool tailored for indie devs and SMEs that lets you drop in a dataset and get a full dashboard with AI-generated insights in just a few minutes.

The Stack & Architecture:

Infrastructure: Google Cloud

AI Engine: Claude. We don't just use LLMs for text summaries; Claude powers the entire data infrastructure layer. It natively reads and understands your schema to power:

Data Studio Assistant: Automatically finds table relationships and generates complex wide views.

Chat-to-Data: Query your datasets naturally without writing SQL.

AI Chart Builder: Describe what you want to visualize, and the AI constructs the exact chart to drop into your dashboard.

Insight Blocks: Reasons directly over the live data to generate executive summaries and strategic takeaways.

The Demo (Why Startups Fail): To stress-test this, I dropped a Kaggle dataset of 409 startup post-mortems into NeoShift. The AI immediately highlighted that 75% of startups are actually killed by established Giants, not a lack of funding. You can see the generated dashboard and read the AI's analysis here: https://bi.neoshift.ai/#/public/dashboard/startup-failure-pr...

Stress-Test My App (The Open Beta Challenge): I just shipped a new "Public Share Links" feature and I want to test the infrastructure limits. I’m extending our Open Beta to March 27th and running a data storytelling challenge.

If you want to break my app or just visualize some cool data, grab a weird dataset from Kaggle, drop it into NeoShift, and share the link. I'm giving a Free Lifetime Enterprise Account (100GB storage, 5 seats) to the dashboard with the best combined score (unique views + feature usage).

To track the competition, I built a live Leaderboard using NeoShift BI itself (connected to the Google Analytics API). You can check the exact rules and current standings here: https://bi.neoshift.ai/#/public/dashboard/competition-leader...

You can unlock a free share link to test it out with code BETA-320430B4 here: https://bi.neoshift.ai/#/register

I’d love your brutal, unfiltered feedback on the UI, the depth of the Claude integration, and how the system handles whatever weird CSVs you throw at it. Happy to answer any questions about the architecture!

Comments URL: https://news.ycombinator.com/item?id=47146229

Points: 1

# Comments: 0

Article URL: https://stabilize.rodmena.ai

Comments URL: https://news.ycombinator.com/item?id=47146222

Points: 1

# Comments: 0