Hacker News: dr_kiszonka

New comment by dr_kiszonka in "CHOMPI portable sampler instrument is now open-source (hardware and software)"

dr_kiszonka — Fri, 02 Oct 2026 17:48:58 +0000

Looks good but consider referencing CHOMPI clearly and sufficiently prominently on your page, not just in the repo.

New comment by dr_kiszonka in "Ask HN: Any nerds out there who've read a lot of research papers?"

dr_kiszonka — Sat, 26 Sep 2026 14:56:17 +0000

Contributors: GCSS had the original idea. JPP tried to talk him out of it. JPP did the first literature search but GCSS lost it. GCSS drafted the manuscript but JPP deleted all the best jokes. GCSS is the guarantor, and JPP says it serves him right.

New comment by dr_kiszonka in "Claude Opus 5.5 Intelligence, Performance and Price Analysis (Max)"

dr_kiszonka — Wed, 23 Sep 2026 00:53:42 +0000

But the chat version does change over time, correct? It has been my experience that Sol's performance has deteriorated significantly.

New comment by dr_kiszonka in "Xiaomi Mimo 2.6 live post-training dashboard"

dr_kiszonka — Thu, 17 Sep 2026 01:17:43 +0000

Very curious that everyone here (so far) seems to assume this dashboard presents real data.

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

dr_kiszonka — Mon, 14 Sep 2026 03:21:40 +0000

It looks fun! Reminds me a bit of TLDRAW computer.

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

dr_kiszonka — Mon, 14 Sep 2026 00:51:46 +0000

Same here. 5.5 would sometimes oneshot relatively complex models even without MCP. Opus, however, even with Fusion 360 MCP access would usually make comically bad models. It seemed that Opus tried to reason through the problem while GPT would come up with a general plan of using whatever tools were available to solve the problem. In chat, it would use the available Python and trimesh. It was also the only time I had ChatGPT work on something for over 30 min without giving up.

New comment by dr_kiszonka in "Google Antigravity introduces Boost deep reasoning (/boost)"

dr_kiszonka — Sat, 12 Sep 2026 19:16:27 +0000

Thanks!

New comment by dr_kiszonka in "λ Snap – An inviting programming language for kids and adults for CS study"

dr_kiszonka — Sat, 12 Sep 2026 08:48:05 +0000

Thank you for your work on Snap! and for finding the time to respond to our comments.

New comment by dr_kiszonka in "Λ Snap – An inviting programming language for kids and adults for CS study"

dr_kiszonka — Fri, 11 Sep 2026 19:13:00 +0000

What device did you get your nephew? When I was looking for one, those listed on the Make Code website were pretty bad and expensive (or unavailable). I would love to find an affordable retro handheld like Miyoo or Anbernic with Make Code support.

New comment by dr_kiszonka in "Show HN: Hacker News, without AI"

dr_kiszonka — Fri, 11 Sep 2026 17:53:18 +0000

This is much better than I anticipated. I don't mind AI that much, but this version of HN surfaces really interesting articles I've missed.

λ Snap – An inviting programming language for kids and adults for CS study

dr_kiszonka — Fri, 11 Sep 2026 17:36:11 +0000

Article URL: https://snap.berkeley.edu/

Comments URL: https://news.ycombinator.com/item?id=49662214

Points: 184

# Comments: 109

New comment by dr_kiszonka in "Logo Programming Language"

dr_kiszonka — Fri, 11 Sep 2026 17:31:18 +0000

So I know a Lisp dialect? Fascinating.

New comment by dr_kiszonka in "YuE2 · Frontier Music with Symbolic Planning"

dr_kiszonka — Fri, 11 Sep 2026 01:48:48 +0000

The vocals in the few songs I listened to are so very clearly AI. I find Suno's much more credible in many genres.

If anyone has tried Suno's Studio 2, I would love to hear your take on it. At $24-30/month, it's approaching Ableton Live 12 Suite rent-to-own territory, without the "own" part.

New comment by dr_kiszonka in "Qwen 3.8 follows GPT-5.5 Pro reasoning prefills"

dr_kiszonka — Thu, 10 Sep 2026 03:53:01 +0000

Could anyone explain to me the difference between thinking traces ("intermediate tokens") and the final responses? Specifically, why is it that Claude Opus 5's reasoning in Code is very easy to follow and sounds quite natural, while its answers are full of these very annoying AI-isms and sentence fragments that are void of meaning?

Are thinking traces and final answers trained for different objectives?

New comment by dr_kiszonka in "Discovery of a new OpenAI agent message board"

dr_kiszonka — Sat, 05 Sep 2026 15:57:18 +0000

I spotted Claude doing it when accessing PubMed.

New comment by dr_kiszonka in "Gemini 3.8 Flash and 3.8 Flash Cyber"

dr_kiszonka — Thu, 03 Sep 2026 01:20:15 +0000

We also have sanctions and incentives to induce specific behaviors, but they don't apply to agents. We can put a muzzle on Guile 3.8 but we can't turn it into Genuine 3..8.

New comment by dr_kiszonka in "Visualizing Combos in Judo in R"

dr_kiszonka — Tue, 01 Sep 2026 17:57:19 +0000

Nice! And it continues the lineage of https://en.wikipedia.org/wiki/Zachary%27s_karate_club

New comment by dr_kiszonka in "AnkiDroid: Google Play no longer allowing Open Collective donation link"

dr_kiszonka — Tue, 01 Sep 2026 17:51:47 +0000

old.reddit now requires a login :/

New comment by dr_kiszonka in "Google Antigravity introduces Boost deep reasoning (/boost)"

dr_kiszonka — Tue, 01 Sep 2026 04:16:58 +0000

I gave up on Antigravity when Google changed its rules around paid plans AI quotas, leading to my AI Pro getting depleted in up to four rounds.

Have they ever gotten around to fixing it or is Antigravity only feasible with their Ultra plan?

New comment by dr_kiszonka in "Show HN: A techno machine in one HTML file, with verifiable renders"

dr_kiszonka — Tue, 25 Aug 2026 21:27:36 +0000

"Well, actually," it is usually not just HTML, but I am with you on liking these single-file apps. Security is complicated because these pages often rely on importing online packages, which can get poisoned in the blink of an eye.