Hacker News: proxysna

New comment by proxysna in "Alpine Linux 3.24.0 Released"

proxysna — Tue, 09 Jun 2026 22:25:00 +0000

Alpine is my go-to nowadays for everything in my homelab except desktop (I use Void btw), because of how dirty the setup to make GPU's work with musl kernel.

New comment by proxysna in "The Cypherpunk Library"

proxysna — Mon, 08 Jun 2026 09:12:02 +0000

Looks really nice, but 10 fps in Firefox.

New comment by proxysna in "Nvidia is proposing a beast of a CPU system for Windows PCs"

proxysna — Sat, 06 Jun 2026 17:44:56 +0000

It's just a DGX spark with faster memory and a windows boot?

New comment by proxysna in "Show HN: Zot – Yet another coding agent harness"

proxysna — Fri, 29 May 2026 22:04:54 +0000

There is a docker registry under the same name. https://zotregistry.dev

New comment by proxysna in "How Shamir's Secret Sharing Works"

proxysna — Tue, 26 May 2026 06:55:13 +0000

I think hashicorp still have an implementation for vaults seal/unseal process. Unless something changed ofc

New comment by proxysna in "Microsoft starts canceling Claude Code licenses"

proxysna — Fri, 22 May 2026 18:02:06 +0000

Feels about right.

I've launched an internal demo of Claude Code and Deepseek on the same day and we burned through our monthly allowance for Claude in just over a week, with more than a half of that budget being spent in one day. With DS people are unable to go through that same amount of money in a month, not even close.

With that Claude feels like an expensive toy, while DS is a shovel, purely because developers do not feel like they are eating into a precious resource while using it. Also it does not feel like there is much of a difference in capability between Claude and DS-pro. DS-pro and flash do feel like sonnet/opus and haiku, but flash is still very-very capable.

New comment by proxysna in "AI didn't delete your database, you did"

proxysna — Tue, 05 May 2026 14:48:15 +0000

“Expert” that does not know what a Terraform is. lol, lmao even

New comment by proxysna in "Running local LLMs offline on a ten-hour flight"

proxysna — Mon, 27 Apr 2026 20:55:58 +0000

Cool, i am mostly a plumber for these things, but do you have any sort of reading that i can go through to wrap my head around it to some degree?

New comment by proxysna in "Running local LLMs offline on a ten-hour flight"

proxysna — Mon, 27 Apr 2026 16:25:17 +0000

I've used only Qwen3.5 so far for work and was, after initial struggles, successful with GPU setup, no mlx. Ngl the fact that they are using `presence_penalty: 0` and no `max_tokens` is weird after that exact setup caused me "initial struggles", but i've set up a simple docker-compose with vllm and qwen3.6 right now to test it out and it worked perfectly fine for me.

Gist with the compose and example of an output. https://gist.github.com/meaty-popsicle/f883f4a118ff345b430c3...

New comment by proxysna in "Running local LLMs offline on a ten-hour flight"

proxysna — Mon, 27 Apr 2026 15:05:11 +0000

You need to set sampling parameters for the llm. Had the same issue with Qwen3.5 when i first started. You can grab them off the model card page usually.

From Qwen3.6 page:

Thinking mode for general tasks: temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0

Thinking mode for precise coding tasks (e.g. WebDev): temperature=0.6, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0

Instruct (or non-thinking) mode: temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0

New comment by proxysna in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

proxysna — Thu, 23 Apr 2026 15:45:17 +0000

OpenWebUI with postgres and vllm for inference, searxng for websearch a few other mcp's for tools.

New comment by proxysna in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

proxysna — Wed, 22 Apr 2026 21:52:02 +0000

This is from an example from my Nomad cluster with two a5000's, which is a bit different what i have at work, but it will mostly apply to most modern 24G vram nvidia gpu.

"--tensor-parallel-size", "2" - spread the LLM weights over 2 GPU's available

"--max-model-len", "90000" - I've capped context window from ~256k to 90k. It allows us to have more concurrency and for our use cases it is enough.

"--kv-cache-dtype", "fp8_e4m3", - On an L4 cuts KV cache size in half without a noticeable drop in quality, does not work on a5000, as it has no support for native FP8. Use "auto" to see what works for your gpu or try "tq3" once vllm people merge into the nightly.

"--enable-prefix-caching" - Improves time to first output.

"--speculative-config", "{\"method\":\"qwen3_next_mtp\",\"num_speculative_tokens\":2}", - Speculative mutli-token prediction. Qwen3.5 specific feature. In some cases provides a speedup of up to 40%.

"--language-model-only" - does not load vision encoder. Since we are using just the LLM part of the model. Frees up some VRAM.

New comment by proxysna in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

proxysna — Wed, 22 Apr 2026 21:22:22 +0000

We do make Claude and Mistral available to our developers too. But, like you said, security. I, personally, do not understand how people in tech, put any amount of trust in businesses that are working in such a cutthroat and corrupt environment. But developers want to try new things and it is better to set up reasonable guardrails for when they want to use these thing by setting up a internal gateway and a set of reasonable policies.

And the other thing is that i want people to be able to experiment and get familiar with LLM's without being concerned about security, price or any other factor.

New comment by proxysna in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

proxysna — Wed, 22 Apr 2026 15:40:51 +0000

Qwen3.5-27B with a 4bit quant can be run on a 24G card with no problem. With 2 Nvidia L4 cards and some additional vllm flags, i am serving 10 developers at 20-25tok/sek, off-peak is around 40tok/sek. Developers are ok with that performance, but ofc they requested more GPU's for added throughput.

New comment by proxysna in "Running a Minecraft Server and more on a 1960s UNIVAC Computer"

proxysna — Tue, 21 Apr 2026 13:42:49 +0000

Yeah, my thought exactly, execution lacked, but i do admire the attempt.

New comment by proxysna in "Running a Minecraft Server and more on a 1960s UNIVAC Computer"

proxysna — Tue, 21 Apr 2026 13:05:54 +0000

What a great write up, and a video too! Even though Minecraft stuff ofc was a bit of a bait, it would be interesting see the answer to "Can it run Doom?".

New comment by proxysna in "Nvidia N1 laptop motherboard picture shows 128GB of LPDDR5x"

proxysna — Fri, 10 Apr 2026 20:11:18 +0000

What a load of e-waste. Nightmare.

New comment by proxysna in "Show HN: Docking – Extensible Linux dock in Python"

proxysna — Wed, 08 Apr 2026 12:02:14 +0000

Hehe, yeah i understand, just something that popped up. Great project!

New comment by proxysna in "Show HN: Docking – Extensible Linux dock in Python"

proxysna — Mon, 06 Apr 2026 20:40:59 +0000

heuh https://www.urbandictionary.com/define.php?term=docking

New comment by proxysna in "How to turn anything into a router"

proxysna — Mon, 30 Mar 2026 14:47:49 +0000

Pleasant thing about routers that is is so simple to build one after learning basics of networking and pretty much any OS or distro can act as one. There are obvious choices like OPN\PFSENSE, OpenWRT, DD-WRT, FreshTomato, but literally any PC with a single Ethernet port can act as one. My favorite setup was a laptop running Ubuntu and the whole router setup was in a single netplan file + dnsmasq for DHCP.

Edit: And ofc best cheap device imo is OrangePI R1 LTS and a whatever usb wifi dongle. Came in clutch a few times, such a nice little device.