Hacker News: arizen

New comment by arizen in "The turbulent AI era is here"

arizen — Thu, 27 Aug 2026 20:55:40 +0000

There are also markets with scarce supply, like desired properties in prime locations. Probably some other similarly structured markets, as well.

So there is also a question, what would be happening to these.

New comment by arizen in "GLM-5.3-Flash"

arizen — Wed, 26 Aug 2026 20:35:57 +0000

Probably apples to apples would be to compare z.ai subscription plans vs API pricing

New comment by arizen in "GLM-5.3-Flash"

arizen — Wed, 26 Aug 2026 15:06:33 +0000

Few weeks ago, I wouldn't expect this statement to be true. Accelerate!

New comment by arizen in "DeepSeek-V4-Flash Update"

arizen — Fri, 31 Jul 2026 09:47:12 +0000

Why not also Qwen?

New comment by arizen in "Claude Opus 5"

arizen — Fri, 24 Jul 2026 20:48:33 +0000

Could these complex/hard to read Fable outputs be sign of some kind of industrial level of intelligence, which us humans may have a hard to comprehend, while it may be also hard for machine to use simpler texts to properly outline all nuances and complexities of concepts it output?

New comment by arizen in "Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA"

arizen — Wed, 22 Jul 2026 07:02:23 +0000

Have you tested Qwen3.8-Max yet?

New comment by arizen in "Kimi K3, Qwen 3.8, and Anthropic's (Potential) Unravelling"

arizen — Mon, 20 Jul 2026 20:14:36 +0000

Probably increasing learning rate have some value for these players which is enabled by subsidized usage.

And cost of acquiring learning via users feedback versus additional learning data value for a given company will dictate whether such company will give up or will continue subsidized usage, despite being unprofitable on the paper, but probably really valuable as a long term strategy for GTM and product development and increased learning rates.

Ask HN: Guidelines on GPT 5.6 Sol steering

arizen — Sun, 19 Jul 2026 12:48:14 +0000

I converged on these principles in my work with Sol so far:

Let Sol medium/high propose and implement minimum viable solution without overengineering and overthinking.

Use XHigh/Max/Ultra sparingly for hardest most complex workloads/advisory. (For Ultra I found it works best for me, when it had to control/supervised multiple complex lanes/subagents simultaneously within same high level strategic context).

Curious to hear what solutions on properly steering and guiding Sol variants have y'all found and use, alongside with your most used variants, and these variants you barely touch/don't use.

PS: I'm not talking about Terra and Luna at all, as I found them to be inferior to Sol variants for my workloads (slower, dumber, not that cheaper etc.)

Comments URL: https://news.ycombinator.com/item?id=48967745

Points: 1

# Comments: 0

New comment by arizen in "Kimi K3: Open Frontier Intelligence"

arizen — Thu, 16 Jul 2026 20:01:02 +0000

And DeepSeek V4 Flash + GLM 5.2 is a really good blend of both (fast/cheap DS + more intelligent GLM)

New comment by arizen in "Train sim created by just one person is being called the best ever made"

arizen — Fri, 10 Jul 2026 08:21:17 +0000

Yes, true creativity usually or mostly comes from real constraints, in my experience.

As, if there are no constraints in some specific area, there is no kinda "survival need" to improve there, hence brain is not working as hard/smart/deep as it could.

New comment by arizen in "GPT-5.6"

arizen — Thu, 09 Jul 2026 17:14:03 +0000

"GPT‑5.6 delivers a step change in design judgment. With only high-level direction, GPT‑5.6 creates tasteful, ergonomic, and functional interfaces. Its stronger computer-use capabilities let it inspect and refine the rendered result—not just generate the underlying code or content—so it can catch visual and functional issues and apply finishing touches before handing the work back."

This one is really promising, as it may allow to close major gap with Claude in design/UI skills

New comment by arizen in "Muse Spark 1.1"

arizen — Thu, 09 Jul 2026 16:42:33 +0000

I'm still confused is it available to public via some sort of subscription?

New comment by arizen in "Fable 5 is Back"

arizen — Thu, 02 Jul 2026 11:07:37 +0000

It was not idiotic from capitalism short-term profit gains point of view.

New comment by arizen in "Previewing GPT‑5.6 Sol: a next-generation model"

arizen — Fri, 26 Jun 2026 17:59:39 +0000

Sol Goodman

New comment by arizen in "Previewing GPT‑5.6 Sol: a next-generation model"

arizen — Fri, 26 Jun 2026 17:57:52 +0000

Ditto on GLM 5.2 + DeepSeek V4 Flash combo.

For most important work (complex, cross-domain inquiries etc.), I still rely on Codex GPT 5.5 though.

New comment by arizen in "Openrouter Fusion API"

arizen — Mon, 15 Jun 2026 11:23:58 +0000

Some anecdata on Fusion: I run same query I used for Fable on OR Fusion and results were worse.

It felt, like Fable was able to kinda grasp very deep knowledge/intelligence layers and outline solution not only in agreeable way, but rather it proposed to prioritize solution items, with discarding some of the items, which made a lot of sense to me.

While Fusion felt more like a bit diversified answer of the same class of pre-Fable SOTA models, without touching the depth of knowledge/intelligence layers, which Fable was able to get, in my very limited tests I did, while Fable was accessible.

New comment by arizen in "GLM 5.2 Is Out"

arizen — Mon, 15 Jun 2026 11:17:40 +0000

OpenCode Go is pretty good in my experience too.

I ended up using DeepSeek V4 Flash as main workload model, while keeping DeepSeek V4 Pro and Qwen 3.7 Plus as advisors on system architecture and other advanced matters to guide DS Flash.

I run a simple benchmark on OpenCode Go models while ago, if anyone want to read more: https://arizenai.com/seven-models-judged-each-other/

New comment by arizen in "MiMo Code is now released and open-source"

arizen — Fri, 12 Jun 2026 11:40:02 +0000

opensourcing software may enable leverage of wider network of contributors to given piece of software,hence software can evolve much more quickly and efficiently.

The number of possible chess games exceeds the number of atoms in the universe

arizen — Thu, 11 Jun 2026 07:35:24 +0000

Article URL: https://spacedaily.com/d-the-total-number-of-possible-chess-games-is-so-large-that-it-exceeds-the-number-of-atoms-in-the-observable-universe-by-some-estimates-there-are-more-possible-chess-games-than-there-are-atom/

Comments URL: https://news.ycombinator.com/item?id=48487437

Points: 1

# Comments: 1

New comment by arizen in "Claude Fable 5"

arizen — Tue, 09 Jun 2026 20:53:59 +0000

Probably there will always be frontier surface which frontier model of a given generation would not be able to automate.