Hacker News: RussianCow

New comment by RussianCow in "MiMo v2.6"

RussianCow — Mon, 21 Sep 2026 22:18:44 +0000

Because the end goal is to ban non-US AI companies from being able to do business in the US.

New comment by RussianCow in "Grok 4.7"

RussianCow — Mon, 21 Sep 2026 19:59:01 +0000

For caching, only if you don't specify your preferred providers and let OpenRouter route each request itself. I have stuff like this in my OpenCode config for each model I use and I regularly get ~90-95% cache hit rates.

    "order": ["relace", "coreweave", "novita", "baseten", "together"],
    "allow_fallbacks": false

It still won't be quite as high as you'd get by just using DeepSeek because occasionally a request will fail and you'll get routed to a backup provider with nothing cached, but it's close enough not to matter in most instances.

But I can't argue with the lower off-peak pricing when using DeepSeek directly. The downside is they train their models on your input, which might be a deal-breaker for many users (as it is for me).

New comment by RussianCow in "Astra for Law"

RussianCow — Sat, 19 Sep 2026 03:49:09 +0000

Honestly, I know several devs who do chores around the house or even play video games while AI does the bulk of the heavy lifting. If nobody cares or even realizes, does it matter? (To be clear, they all work remotely.)

New comment by RussianCow in "Nvidia dismisses "circular financing", says every $1 it invests brings back $100"

RussianCow — Sun, 13 Sep 2026 12:35:51 +0000

Which doesn't matter if they're still not profitable.

New comment by RussianCow in "Qwen 3.8 27B available on Cerebras at 1500 tokens/s"

RussianCow — Thu, 03 Sep 2026 21:51:30 +0000

I don't see any kind of input cache discount listed on your pricing page. Do you offer that, or is all input priced the same?

New comment by RussianCow in "Qwen 3.8 27B available on Cerebras at 1500 tokens/s"

RussianCow — Thu, 03 Sep 2026 21:47:39 +0000

The issue is that all input (including context) counts towards that limit. So 10 requests with 50k of context will blow through the limit, even if little to no output was generated, which is incredibly easy to do with agentic workloads.

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

RussianCow — Wed, 02 Sep 2026 15:24:33 +0000

This is pretty terrible advice when there are dozens of AI inference providers out there serving great models with significantly more cost effectiveness than you'd get from buying your own hardware.

New comment by RussianCow in "Can I opt out of my input or output data being used for training?"

RussianCow — Wed, 02 Sep 2026 14:35:40 +0000

I don't trust any company with their word on anything. Luckily, privacy policies are legally binding.

New comment by RussianCow in "Hy4 preview"

RussianCow — Sun, 30 Aug 2026 16:56:50 +0000

Presumably the number that OpenRouter shows is averaged across all requests.

New comment by RussianCow in "Hy4 preview"

RussianCow — Sun, 30 Aug 2026 16:55:46 +0000

Most providers do what's called "prefix caching", where each turn in a session is cached such that sending new messages with the exact same "prefix" (set of previous messages) gives you the cache read price on that input instead of the full price. As long as you're not changing your system prompt, available tools, etc mid-session, you automatically benefit from this.

New comment by RussianCow in "Hy4 preview"

RussianCow — Sun, 30 Aug 2026 16:48:23 +0000

They're not "docking points", they're calculating it in the most straightforward way. If I start a session and the majority of requests are sent to Provider A, and my last request gets routed to Provider B, I have a 0% cache hit rate with Provider B. I'm very curious how else you expect this to be calculated? Do you think they're completely omitting requests that switch providers mid-session?

FWIW, I get significantly higher than listed cache hit rates when I pin my session to a specific provider, which is further evidence of the above.

New comment by RussianCow in "Claude Session URL appended to commit messages and PR descriptions by default"

RussianCow — Sun, 30 Aug 2026 16:43:27 +0000

But presumably everyone in your company/team is using Jira, so it's not an "ad" because it's a product already used internally. Claude is appending these links to all commits by default, whether or not others on the team use Claude. Those are very different things.

New comment by RussianCow in "Hy4 preview"

RussianCow — Sun, 30 Aug 2026 06:45:28 +0000

How else would you expect them to calculate it?

New comment by RussianCow in "Hy4 preview"

RussianCow — Sun, 30 Aug 2026 06:41:19 +0000

This is very much NOT my experience in practice, even though it's how I would expect it to work. OpenRouter will happily bounce you between several providers (none of which have downtime) even within the same session. Requesting specific providers is the only way I've been able to hit a cache rate above 90%.

New comment by RussianCow in "Boot a Virtual iPhone via Apple's Virtualization.framework"

RussianCow — Sat, 29 Aug 2026 02:26:41 +0000

Can you give some examples of where it matters? I'm genuinely curious.

New comment by RussianCow in "Boot a Virtual iPhone via Apple's Virtualization.framework"

RussianCow — Sat, 29 Aug 2026 01:43:17 +0000

That explains the difference, but what's the purpose?

New comment by RussianCow in "GLM-5.3 is now open-weight"

RussianCow — Fri, 28 Aug 2026 23:04:36 +0000

Interesting. I wonder where they're getting their data from then, because they list Z.ai under Singapore, but everything I'm finding says they're based in Beijing. Same with MiniMax.

New comment by RussianCow in "GLM-5.3 is now open-weight"

RussianCow — Fri, 28 Aug 2026 19:08:09 +0000

I don't think that's right, or if it is, OpenRouter has incorrect data. Several Chinese companies (headquartered in China) have Singapore listed as their region on OR. And some companies, like Alibaba Cloud, have multiple regions listed.

I'm happy to be proven wrong, but this makes me think that the region is where the servers are, not where the HQ is.

New comment by RussianCow in "GLM-5.3 is now open-weight"

RussianCow — Fri, 28 Aug 2026 16:19:00 +0000

If you click on the provider name, the panel that pops up shows a "Region" value. Not every provider lists their region, however.

New comment by RussianCow in "GLM-5.3: Frontier coding with emergent cyber capabilities"

RussianCow — Sat, 15 Aug 2026 00:46:44 +0000

You might be joking, but a harness provides much more than just the prompts: at a minimum, it provides the system prompt and the built-in tools that the LLM can use, but it can also provide things like subagent management, custom compaction logic, session forking, etc.