Hacker News: kzrdude

New comment by kzrdude in "DeepSeek V4 Flash 0731"

kzrdude — Sat, 15 Aug 2026 15:05:47 +0000

Is dflash still free there after the 0731 upgrade?

New comment by kzrdude in "Qwen 3.8 27B"

kzrdude — Sat, 15 Aug 2026 09:09:12 +0000

So what's needed to solve this is that someone publicly and popularly benchmarks using the chat template. So that they have an incentive to look good using it.

New comment by kzrdude in "Qwen 3.8 27B"

kzrdude — Fri, 14 Aug 2026 19:57:38 +0000

Is the chat template used at all when they benchmark the model?

New comment by kzrdude in "Qwen 3.8 27B"

kzrdude — Fri, 14 Aug 2026 19:49:12 +0000

So you happen to know how to demonstrate this higher capacity, what do we mean by that? Can understand bigger codebases? And how do we explain it, what part of the model does it?

New comment by kzrdude in "Qwen 3.8 27B"

kzrdude — Fri, 14 Aug 2026 19:46:35 +0000

I think that's called test-time scaling i.e using more tokens at infer time to squeeze out higher model performance. That's must be part of the explanation for good benchmark results.

New comment by kzrdude in "The Indo-European Family Tree"

kzrdude — Thu, 13 Aug 2026 18:35:32 +0000

I just think, look at how much happens in a generation or two right now. No wonder languages can change quickly.

New comment by kzrdude in "The Indo-European Family Tree"

kzrdude — Thu, 13 Aug 2026 18:34:47 +0000

It's believed that basque is a remnant of earlier people before agriculture and indo europeans arrived.

New comment by kzrdude in "Meta Muse Glimmer – open weights 30B local coding model"

kzrdude — Mon, 10 Aug 2026 11:47:53 +0000

Any pointers to more info about that? Sounds interesting.

New comment by kzrdude in "Python string literals are kinda funny"

kzrdude — Sat, 08 Aug 2026 21:31:02 +0000

I find it funny and satisfying that Python has converged to it's own "printf", by which I mean print(f"").

New comment by kzrdude in "Can Intel finally beat ARM on performance per Watt?"

kzrdude — Sat, 08 Aug 2026 21:17:34 +0000

Yes, it seems to be a new process, 18A. (18 Ångström, allegedly). Manufactured in North America. If the shiny video they did didn't fool me, it's a two sided wafer process, power grid on one side, then the silicon wafer (with through-silicon vias), then transistors, then signal layers.

The processor is this one: Intel Core 5 processor 320 https://www.intel.com/content/www/us/en/products/sku/246018/... in the Wildcat Lake series.

New comment by kzrdude in "As a Windows user, it's a surreal way to install a program"

kzrdude — Sat, 08 Aug 2026 17:26:11 +0000

As a mac user (also a linux user), I got used to it and it always seemed better than the alternative (weird install programs).

New comment by kzrdude in "AMD acquires Taalas to boost inference performance by etching models in silicon"

kzrdude — Sat, 08 Aug 2026 11:23:44 +0000

Yes, "this one is obsessed with the golden gate bridge" will be for real this time.

New comment by kzrdude in "DeepSeek V4 Flash 0731"

kzrdude — Fri, 07 Aug 2026 22:09:01 +0000

Just trying to understand, https://opencode.ai/docs/go/#privacy currently says DeepSeek V4 Flash has 0 days data retention.

> DeepSeek V4 Flash: ZDR agreement is renewed monthly. The current agreement is valid through August 31, 2026.

Is there other info I should be aware of w.r.t data retention with opencode go? It's hosted in China, so other middlemen may be active (I doubt it, but possible)?

New comment by kzrdude in "AMD acquires Taalas to boost inference performance by etching models in silicon"

kzrdude — Fri, 07 Aug 2026 05:18:41 +0000

The investment in bigger machines at the fab might set you back billions. I don't know about the lithography technology either, how easy you can scale it to larger wafers?

New comment by kzrdude in "Karpathy’s Pelican"

kzrdude — Mon, 03 Aug 2026 10:14:15 +0000

We don't use these models by way of "one shot" so I don't see why it's relevant. It's clearly useful to let them iterate.

New comment by kzrdude in "Less Coffee, Better Sleep"

kzrdude — Mon, 03 Aug 2026 10:13:07 +0000

People just have different tolerance. Coffee strictly before noon works well for me.

New comment by kzrdude in "Qwen3.8-Max: A New Bar for Coding and Cowork"

kzrdude — Mon, 03 Aug 2026 09:58:56 +0000

It has produced some locally useful models.

New comment by kzrdude in "Ten advances in mathematics and theoretical computer science"

kzrdude — Sat, 01 Aug 2026 12:51:21 +0000

Do mathematicians have the right to say "no AI PRs please, the volume is too much" just like how some open source maintainers do it? I guess they feel a loss of control, there is no way to turn the hose off.

Thinking of this a little bit with the perspective of every new proof as a burden, dumped for review by actual mathematicians.

New comment by kzrdude in "Apple Will 'Watch Everything Burn' When AI Bubble Bursts"

kzrdude — Fri, 31 Jul 2026 21:37:10 +0000

It's summer vacation time.

New comment by kzrdude in "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis"

kzrdude — Fri, 31 Jul 2026 14:20:39 +0000

The thinking trace was (preview) frustrating to read, I think I'd prefer a summary view of it at this point.