Hacker News: ponyous

New comment by ponyous in "Using LLM for confidence scoring is a bad idea"

ponyous — Thu, 22 Oct 2026 09:27:58 +0000

That's very well put. I noticed the issue myself when building my own evals and naturally started drifting into the direction you are describing, but you describe it so succinctly that now I am seeing how it should actually be done. Thank you

New comment by ponyous in "WorldTUI: I trained a 360M model to pretend to be a terminal"

ponyous — Thu, 22 Oct 2026 09:01:32 +0000

One of the first very cool demos I've seen of ChatGPT was in a post where the author prompted AI "pretend to be a terminal". Then went on to execute commands like `cd .fbi`, `ls files`, `cat files/confidential.txt`. Very fun demo. Looking forward to exploring similar idea with this terminal

New comment by ponyous in "Ask HN: What do you hate the most about GitHub Actions?"

ponyous — Tue, 20 Oct 2026 16:14:55 +0000

Organization. I am always lost. Gitlab pipelines seem much more straigh forward to understand for me.

New comment by ponyous in "OpenSCAD the Programmers Solid 3D CAD Modeller"

ponyous — Sun, 11 Oct 2026 01:06:34 +0000

As someone who’s been building AI for CAD[0] with OpenSCAD. I’m very thankful to OpenSCAD maintainers it’s been a life saver. Terse syntax, a lot of plugins and great CSG kernel.

People keep saying I should use other kernels for fillets and stuff, but from my testing AI is just not as good with, for example build123d as it is with OpenSCAD.

[0]: grandpacad.com

New comment by ponyous in "LLM Pareto Frontier over time [video]"

ponyous — Sun, 11 Oct 2026 00:41:24 +0000

Was curious how the race looks like if you look at Pareto frontier of LLM capability since this hype started. So I asked Claude to generate me a video. And it's quite insane how good it is and I thought you guys might find it interesting too.

LLM Pareto Frontier over time [video]

ponyous — Sun, 11 Oct 2026 00:41:24 +0000

Article URL: https://www.youtube.com/watch?v=I01v36VIrno

Comments URL: https://news.ycombinator.com/item?id=50038599

Points: 5

# Comments: 1

New comment by ponyous in "Ask HN: What are people doing with their OpenClaw set ups now?"

ponyous — Mon, 28 Sep 2026 13:53:19 +0000

How did you connect it to search console and do you like the integration? I would like to do this too, and eventually to Google Ads too.

New comment by ponyous in "Cartesian – AI 3D Modeling for Design"

ponyous — Tue, 15 Sep 2026 23:00:26 +0000

It really depends on the complexity of the model you are trying to generate and the budget you have.

New comment by ponyous in "Steam Frame starts at $1059"

ponyous — Tue, 15 Sep 2026 07:51:37 +0000

You mean people blind in one eye? I think it still works, but obviously depth perception will not be there

New comment by ponyous in "Ask HN: What are you working on? (September 2026)"

ponyous — Mon, 14 Sep 2026 06:54:43 +0000

Surprisingly Gemini models. Spatial understanding seems to be the best for the price and speed.

New comment by ponyous in "Ask HN: What are you working on? (September 2026)"

ponyous — Sun, 13 Sep 2026 18:00:05 +0000

https://grandpacad.com - AI modeling for 3D printing Been at it for about year and a half.

Really exciting stuff is happening literally every month, because underlying models are getting better and better. When I started it was pretty basic: “make a cube with a hole through it”. Now it’s at the point “make a raspberry pi 4 case” and the agent searches, builds, verifies…

What surprised me in this process is how little meaning AI benchmarks have. Pareto frontier for my use case looks completely different than any other benchmark portrays.

New comment by ponyous in "DeepSeek v4.1 Flash"

ponyous — Thu, 10 Sep 2026 12:23:01 +0000

> I have a feeling that the price of 1 million tokens transmitted over the internet is more expensive than cache hit.

And this kinda makes sense. What is cheaper few KB of disk space or internet bandwidth?

New comment by ponyous in "Claude Fable 5.1 and Claude Mythos 5.1"

ponyous — Wed, 02 Sep 2026 09:32:21 +0000

We went for 16% intelligence bump according to artificial analysis for +82% of the cost. Interesting.

Comparing 4.8 Opus with Fable 5.1

New comment by ponyous in "Gemini 3.7 Flash"

ponyous — Thu, 13 Aug 2026 18:16:16 +0000

On my benchmark where AIs generate ~20 different 3D models about 1/2 the time of Opus and 1/3 of the time of Kimi K3 and 2/3 of time of sonnet.

New comment by ponyous in "Ask HN: What are you working on? (August 2026)"

ponyous — Mon, 10 Aug 2026 09:26:05 +0000

Yeah you can absolutely do multiple parts, although the mating features are still a bit rough you can do a lot already.

New comment by ponyous in "Ask HN: What are you working on? (August 2026)"

ponyous — Sun, 09 Aug 2026 19:19:58 +0000

GrandpaCAD - AI 3D modeling software focused on simplicity. Made it so even my grandpa could model. He’s been asking me for years when will I teach him how to 3D model. I tried, we failed and then I seen him use ChatGPT so I knew there was a better way than traditional CAD tools.

Recently we also got European funding and the project got some traction. Very exciting times ahead.

https://grandpacad.com

New comment by ponyous in "DeepSeek V4 Flash 0731"

ponyous — Fri, 07 Aug 2026 21:50:47 +0000

You are right, relatively to other llm providers this is not slow. But if you think what is possible when you have 1000t/s a sec you might find it slow.

New comment by ponyous in "Modern email can be built from borrowed parts"

ponyous — Mon, 27 Jul 2026 17:06:27 +0000

Spark email client supports this. It’s great

New comment by ponyous in "Are AI labs pelicanmaxxing?"

ponyous — Thu, 23 Jul 2026 12:54:14 +0000

I agree with the conclusion and am happy to see this blog post, but this killed a bit of credibility for me:

> Using a single LLM judge for scoring. Every score here comes from one model, GPT-5.6 Luna, looking at one image at a time. I didn’t do much alignment and didn’t check how often it agrees with itself on a re-run.

Having used a similar setup (with previous gen LLMs) to evaluate the 3D models that my product[0] generates, it turned out there was no correlation at all. LLM judgments were very much random and I assume judging SVGs is not that far from judging 3D models. I guess I have to re-test this with current gen.

[0]: https://grandpacad.com

New comment by ponyous in "Hy3"

ponyous — Fri, 10 Jul 2026 10:36:16 +0000

Interesting way to show off a model last on every benchmark. Not sure any other lab is doing this