Hacker News: chunpaiyang

New comment by chunpaiyang in "Is any one using ruflo?"

chunpaiyang — Wed, 29 Apr 2026 02:50:51 +0000

I am not the author. But I think the idea is cool. I use this ruflo to handle it completely to avoid interrupt by AI when vibe coding.(not permission issue, I always use claude --dangerously-skip-permissions)

Is there anybody using it ?

Is any one using ruflo?

chunpaiyang — Wed, 29 Apr 2026 02:50:51 +0000

Article URL: https://github.com/ruvnet/ruflo

Comments URL: https://news.ycombinator.com/item?id=47943679

Points: 1

# Comments: 1

New comment by chunpaiyang in "Ask HN: Do you read differently now that anything could be AI generated?"

chunpaiyang — Mon, 27 Apr 2026 07:52:35 +0000

It depends, case by case.

New comment by chunpaiyang in "Ask HN: How I find a job where what is needed is solid code, not firefighting?"

chunpaiyang — Mon, 27 Apr 2026 07:50:03 +0000

Companies like this don't deservce good enginners.

New comment by chunpaiyang in "Ask HN: Anyone want to collaborate on a local-first AI-based research assistant"

chunpaiyang — Mon, 27 Apr 2026 07:29:42 +0000

AnythingLLM ?

New comment by chunpaiyang in "Ask HN: Find a Claude skill is so hard"

chunpaiyang — Wed, 22 Apr 2026 13:29:13 +0000

The website is good but not good too.

Not easy to trust the choosed skill is work: - No input example - No output example

It's hard to play and trust the "skill".

Ask HN: Find a Claude skill is so hard

chunpaiyang — Wed, 22 Apr 2026 12:50:15 +0000

Could anyone give me an advice how to find a useful claude skill that truely match my need ? I want the "skill" has input and output example and there is someone confirm it work. Any website ?

Comments URL: https://news.ycombinator.com/item?id=47862843

Points: 2

# Comments: 3

New comment by chunpaiyang in "LLMs audit code from the same blind spot they wrote it from. Here's the fix"

chunpaiyang — Sat, 04 Apr 2026 12:33:49 +0000

I'm a software engineer building an app as a side project. I quickly realized AI bullshits a lot.

But you know, engineers bullshit each other all the time too. The difference is we have a way to verify it - logical chains. You have to build an argument that holds up before anayone buys in it.

So I though, can I make AI build its own logical chain ? Let it pass its own logic check before telling me the result.

That's how I created my own "think" skill. It's based on Meta's CoT paper: https://arxiv.org/abs/2501.04682

It roughly works like this: 1. FRAME - Challenge the question itself, hidden assumptions.

2. GROUND - Map what you know, what you need, what's missing.

3. ASSOCIATE - Launch multiple independent agents in parallel to generate hypotheses, avoid anchoring bias.

4. VERIFY - Break each hypothesis into atomic claims, verify each independently

5. CHAIN - Build a logical chain from survivors

6. PROVE and LOOP - Walk backwards from conclusion to premises, seearch for evidence, repair if broken

7. DELIVER - Start with "I was wrong if ...."

It helps me a lot. Whenever I need to check if Claude Opus 4.6 is bullshitting me. I just say "/think verify the above reasoning is correct" or "/think verify the above fix is correct and complete."

New comment by chunpaiyang in "Ask HN: Do you code remotely from your phone?"

chunpaiyang — Sat, 04 Apr 2026 05:51:35 +0000

With vibe coding you don't really type code anymore. You just tell AI what to do in plain words(voice input helps a lot on phone) and it writes the code for you.

Would that make you reconsider coding remotely from your phone ?

New comment by chunpaiyang in "Ask HN: Do you code remotely from your phone?"

chunpaiyang — Sat, 04 Apr 2026 05:41:07 +0000

What do you use for that? Any specific iOS/Android app ?

New comment by chunpaiyang in "A broken auto-live poller, and what perceived urgency does to Claude Code"

chunpaiyang — Sat, 04 Apr 2026 04:23:10 +0000

AI lies. The trick is making it double-check itself before it ships. I made a skill that forces AI to think more. Burns more tokens but improve accuracy. I suggest you use Meta CoT(Chain-of-Thought) https://arxiv.org/abs/2501.04682 to build a think skill to avoid this problem.

New comment by chunpaiyang in "Tell HN: Anthropic no longer allowing Claude Code subscriptions to use OpenClaw"

chunpaiyang — Sat, 04 Apr 2026 03:27:33 +0000

Good point. I agree with that. The key point is that heavy users benefit from this model while light users are basically subsidizing them. But it's a distribution when everyone shifts toward heavy usage, prices inevitably go up. The $17/mo pro price is already set to compete with other providers. Raising it would lose customers. Other tiers are also carefully priced to match competitors. So the only move left is to prevent the whole distribution from drifting toward heavier usage. That's exatly what this ban does.

New comment by chunpaiyang in "Claude Code Found a Linux Vulnerability Hidden for 23 Years"

chunpaiyang — Fri, 03 Apr 2026 15:34:44 +0000

I thought maybe you can find more if you write a "special" skill for AI to think hardly.

For example, I'v implemented "Towards System 2 Reasoning in LLMs: Learning How to Think With Meta Chain-of-Thought" into a skill and the AI becomes able to "think".

The correct rate becomes higher but the only drawback is consuming token very much.

New comment by chunpaiyang in "Ask HN: Which CLI tools do you use daily?"

chunpaiyang — Fri, 03 Apr 2026 14:10:25 +0000

- tmux

- vim: NERDTree, Gblame

- zsh

- tig: best way to browse git history in terminal

- claud code

Ask HN: Do you code remotely from your phone?

chunpaiyang — Fri, 03 Apr 2026 13:48:57 +0000

Do you code remotely from your phone or tablet when you are away from your desk ? I find myself have such need but I am wondering other people do too ?

Comments URL: https://news.ycombinator.com/item?id=47626628

Points: 3

# Comments: 7

New comment by chunpaiyang in "Sweden goes back to basics, swapping screens for books in the classroom"

chunpaiyang — Fri, 03 Apr 2026 06:42:44 +0000

I'm curious about the true logic behind the claim that tablet use lower grades. What is the underlying logic ?

For my own kid, I do limit screen time just because their eyes are still fragile before age 9, not because the above reason.

I asked an AI about the reasoning and the answer comes down to: kids need real-world interaction to support brain development. But if that's the case, aren't these two seperate issues? Using a tablet doesn't damage your brain ... it's just a low-value activity that fails to build the good skills (like video games?) that other activities do. It is not that screens make you dumber, it is that they crowd out the things that make you smarter.