Hacker News

Subscribe to Hacker News feed
Hacker News RSS
Updated: 17 min 49 sec ago

MLX Vulkan Back End

Thu, 05/21/2026 - 9:43am

Article URL: https://github.com/goniz/mlx-vulkan

Comments URL: https://news.ycombinator.com/item?id=48222428

Points: 2

# Comments: 0

Categories: Hacker News

Can AI solve this Bongard problem?

Thu, 05/21/2026 - 8:54am

Article URL: https://imgur.com/a/idyH0Kh

Comments URL: https://news.ycombinator.com/item?id=48221864

Points: 1

# Comments: 1

Categories: Hacker News

Microsoft Build CLI

Thu, 05/21/2026 - 8:51am
Categories: Hacker News

Show HN: I Made a Claude Skill for SDD

Thu, 05/21/2026 - 8:49am

At my work they provided a single Claude subscription for everyone on the team. To be honest I like kiro better as it provides a way better SDD management. But the company can't provide it and I can't afford it yet. Turns out I had the skill creator skill in my claude instance so I made use of it to create this Skill. I made it fully by using Claude but I wanted to make it open source, so I asked it to help me make tests and preparations for it, even a CI to run python tests. Well, we got this results with it: - Phase 2A: 67 static assertions (Python script, runs in CI) - Phase 2B: 15 behavioral tests (live Claude Code session) - Phase 2C: 53 generation quality checks across 3 end-to-end flows All of these passed and the CI also passed (after a few tries). I made it to suit my way of prompting and coding and based it off kiro's SDD management, but I want it to be publicly available and used by many people. According to claude some of the testers need to fit the following criteria: 1. Developer starting a real new project from scratch 2. Solo dev with an active side project (greenfield or partial codebase) 3. Team lead whose team uses multiple AI tools 4. Developer with an existing codebase and no written specs 5. Developer who actively uses 3+ AI coding tools It's actually a blind test, no guiding, just try it if you can, I'd really appreciate your help. The repo is here: https://github.com/FredAntB/Spec-Driven-Development

Comments URL: https://news.ycombinator.com/item?id=48221805

Points: 1

# Comments: 0

Categories: Hacker News

Kubernetes in Anger

Thu, 05/21/2026 - 8:47am
Categories: Hacker News

Ask HN: Does anyone believe role-play AI is effective for training?

Thu, 05/21/2026 - 8:46am

We built Socratize, an AI-based training tool where employees practice real workplace scenarios instead of watching videos or taking quizzes.

Most corporate training is passive. People watch content, click through slides, pass a quiz, and forget most of it within days.

We wanted to try a different approach: learning through dialogue and repetition.

With Socratize, users enter realistic scenarios and have to respond to an AI “counterparty”.

For example:

A sales rep practices handling: “Your product is too expensive” A support agent practices de-escalating an angry customer An employee practices explaining a compliance rule in their own words

The AI responds like a realistic counterpart, challenges weak arguments, and continues the conversation until the user improves or fails the scenario.

The goal is simple: replace passive training with active practice.

We’re using Claude to generate responses and evaluate the quality of the user’s arguments based on context and reasoning. Each session is stored so teams can see where people struggle and what scenarios are hardest.

Tech stack:

Next.js (frontend) Node.js backend PostgreSQL Claude API

We launched the MVP last week.

We’re still early and would really appreciate feedback from people here:

Does this approach actually make sense for training? What use cases are we missing? What would make this fail in real companies? Is this fundamentally useful or just “interesting but not needed”?

Try it here: https://socratize.io

Free tier, no credit card required.

Happy to answer any questions.

Comments URL: https://news.ycombinator.com/item?id=48221788

Points: 1

# Comments: 0

Categories: Hacker News

Pages