Hacker News: vcryan

New comment by vcryan in "Inkling: Our Open-Weights Model"

vcryan — Wed, 15 Jul 2026 21:00:37 +0000

I'm surprised that Nemotron gets mentioned at all. In my experiments with it for coding tasks it performed extremely poorly, essentially unusable.

New comment by vcryan in "Samsung Health app threatens data deletion if users opt out AI training"

vcryan — Mon, 13 Jul 2026 21:30:50 +0000

Yes, please - delete my health data. I want my health data - I didn't want Samsung or anyone else to have it unless I provide it. And even then, you can't keep it - you can look at it. It's mine.

New comment by vcryan in "Ask HN: Is anyone experimenting with different ways of using LLMs for coding?"

vcryan — Sat, 04 Jul 2026 04:01:18 +0000

I do all my development as a longish running pipeline that mimics how I used to write software.

There is a high-level plan, it gets decomposed into steps that get performed sequentially. If during a step something occurs that challenges the original assumptions, the remaining steps can be reconfigured or the implementation stopped.

Well, I guess it's more rigorous than me, because it does a lot of quality checking along the way and if course, in the end.

This is the most boring approach to this work as you can imagine. But, after running this process hundreds of times and tuning it, today I can do about 10-20 PRs a day that are often quite good. I manually review and manually test and as models and my pipeline have improved the % of PRs that require zero human intervention after the initial planning is getting higher. Maybe 50%. For the other half, it's good, but there are often glaring bugs, especially for UI/UX stuff.

New comment by vcryan in "Asian AI startups launch Mythos-like models"

vcryan — Sat, 27 Jun 2026 14:42:44 +0000

We are all people. This ally-of-the-west framing is propaganda. Who has harmed me more: this US or China? Who do I have more in common with: a tech worker in China or a US government official?

(I'm based in US - I use the best tech for the task).

New comment by vcryan in "Beer CSS – Build material design in record time"

vcryan — Sat, 27 Jun 2026 14:39:04 +0000

I'm with the herd on this one: unless you're looking for a sort of "retro" feel, most designers are not reaching for Material Design.

New comment by vcryan in "Show HN: Oak – Git alternative designed for agents"

vcryan — Mon, 22 Jun 2026 20:51:05 +0000

I think git it the perfect system to bring order AI coding. Generate code however you like, but ultimately, it goes through a sensible, proven, and auditable pipeline. As someone also building AI tools, git it something I find myself "building with" - not against.

New comment by vcryan in "I built Ponytrail, a local audit trail for AI coding-agent edits"

vcryan — Mon, 22 Jun 2026 15:13:31 +0000

My entire repo is an audit trail of AI coding-agent edits

New comment by vcryan in "Apertus – Open Foundation Model for Sovereign AI"

vcryan — Mon, 22 Jun 2026 00:36:06 +0000

Maybe I'll give Nemotron another try. Yesterday I used the latest one on OpenRouter and it was bad - worse than StepFun

New comment by vcryan in "Prefer duplication over the wrong abstraction (2016)"

vcryan — Sun, 21 Jun 2026 18:53:52 +0000

The sweet spot is really duplicating the wrong abstraction: I see you Claude!

New comment by vcryan in "GLM 5.2 Is Out"

vcryan — Sun, 14 Jun 2026 00:54:51 +0000

I used to use GLM before I knew about coding subscriptions and it was okay. I've tried every version since 4.6 and this one is doing a great job a spec-implementation runner. If I had to guess... somewhere between Sonnet and Opus in terms of quality. Z.ai's issue has been service reliability. So far so good on day one.

New comment by vcryan in "The desperation of NYTimes"

vcryan — Thu, 04 Jun 2026 23:30:10 +0000

There is plenty of biased and misleading information on the Internet, I can't imagine paying the NYT to get it from them.

New comment by vcryan in "AI, Ashby Engineering, and the future"

vcryan — Thu, 04 Jun 2026 23:27:46 +0000

The irony is that you can also use AI coding to just build your own Ashby since building Ashby is so easy to do with AI (apparently).

New comment by vcryan in "MAI-Thinking-1"

vcryan — Tue, 02 Jun 2026 20:59:29 +0000

It really looks like they used Claude to design this webpage. I guess the color taupe it the marker of good AI today.

New comment by vcryan in "Bun support is now limited and deprecated"

vcryan — Sat, 23 May 2026 00:14:20 +0000

I don't think it matters how code is produced -- it matters what it achieves. Is there evidence that there is something wrong with recent Bun releases?

New comment by vcryan in "Israel's AI targeting system: how data from a phone become a death sentence"

vcryan — Sun, 10 May 2026 14:24:13 +0000

This is terrorism

New comment by vcryan in "Palantir employees are starting to wonder if they're the bad guys"

vcryan — Thu, 23 Apr 2026 20:19:14 +0000

Reminds me of the day I realized that, during my lifetime, my country, the US, caused the death of 1M Iraqis -- for no apparent reason.

New comment by vcryan in "Copilot edited an ad into my PR"

vcryan — Mon, 30 Mar 2026 06:04:25 +0000

I'm not a fan of LLM's injecting themselves into PR/commit content. If you use multiple models, basically whichever one is operating git gets all the credit. But, even if you wrote all the code yourself, and just submitted the PR with Claude Code (or whatever) it would attempt to take credit for the changes.

I currently have rules in all of my skill files forbidding models from advertising themselves or taking credit.

New comment by vcryan in "Iran-linked hackers breach FBI director's personal email"

vcryan — Sat, 28 Mar 2026 14:53:26 +0000

This is one of the risks of dating a Mossad agent.

New comment by vcryan in "Ask HN: How do you deal with people who trust LLMs?"

vcryan — Thu, 19 Mar 2026 02:30:37 +0000

The people who trust bad information from LLMs are the same people who trusted bad information from search results and new articles, it just takes them less time to get bad information.

New comment by vcryan in "Ask HN: How is AI-assisted coding going for you professionally?"

vcryan — Sun, 15 Mar 2026 20:18:01 +0000

Love it. I can create closer to the speed of decision making.