Hacker Newsnew | past | comments | ask | show | jobs | submit | dstryr's commentslogin

i have this exact same machine and 3.8-27b runs well, albeit slow at about 6 t/s.

i have it complete tasks in the background as i work on other things.

for comparison, my 3090 runs it at about 60 t/s.


What are you using to run the model and as coding harness?


llama.cpp + pi.dev


This is not my experience at all. Even the Nous Research guys have stated that "Qwen3.6-27B is the canonical local model to use Hermes Agent with" [https://old.reddit.com/r/LocalLLaMA/comments/1sz2y76/ama_wit...]. I am finding the same when used with Pi and OpenCode.

Gemma will just stop mid-tool call. It's been slower and I've had to reduce context size to run it. Qwen3.6 27b has been rock solid using club 3090's single card setup for agentic use -- https://github.com/noonghunna/club-3090/blob/master/docs/SIN...


I'm talking about automation generally, not agent loops.

E.g. prompt A to achieve X, output in format Y. Use Y to do something in prompt B.

Agentic loops will underperform deterministic control flow pipelines (with non-determinism constrained to LLM calls).

Agents are more general, which is the main advantage. But inherently a more general solution will waste context on unnecessary reasoning.

Try asking the smaller Qwen models to output a JSON in a specific format. It basically can't do it consistently with a moderately sized prompt unless you constrain the token generation via GGML or are extremely repetitive and specific about it. (Thinking disabled)

Gemma 4 will do it correctly pretty much 100% of the time. (Thinking disabled)

Applies to other rule following as well in my experience.

Qwen may be better at toolcalling and certainly probably codegen.

It seems to me Google explicitly designed Gemma for edge device automation, and didn't fine tune for agentic or coding use cases.


I would purchase [2] used 3090's as close to $600 as you can. The 3090 still remains the price-performance king.


Give this project a try. I've been using it with promising results.

https://github.com/matthsena/AlcheMark


I tried with one PDF and was surprised to see it connect to some cloud service:

  2025-05-14 07:58:49,373 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): openaipublic.blob.core.windows.net:443
  2025-05-14 07:58:50,446 - urllib3.connectionpool - DEBUG - https://openaipublic.blob.core.windows.net:443 "GET /encodings/o200k_base.tiktoken HTTP/1.1" 200 361 3922
The project's README doesn't mention that anywhere...


The project's README mentions that it uses tiktoken[0], which is a separate project created by OpenAI.

tiktoken downloads token models the first time you use them, but it does not mention that. It does cache the models, so you shouldn't see more of those connections, if I'm understanding the code correctly.

[0] <https://github.com/openai/tiktoken>


I'll check it out!


Not that unusual in the context of trying to outshine anything that could be released tomorrow at llamacon.


If you want a dick move like this it's better to do so after. OpenAI consistently pull this trick on Google.


I built a site that saves time by summarizing YouTube videos or news articles by simply inputting the URL. The tool preserves the original context, allowing users to ask follow-up questions.

I'd like to continue building fun projects like this until I find a market. I work in Phase 1 clinical trials and the end goal would be to implement some of these efficiencies into health technologies.

https://tldw.pw/


- Install Shelter from F-Droid

- Install the Aurora store apk inside of Shelter

- Open Aurora store in Shelter's work profile and use like a normal play store and all apps installed within Aurora remain sandboxed


There is also a built-in Work Profile feature now, under Settings -> System -> Multiple Users.

It basically does the same thing under the hood.


Not mentioned here, and more importantly in my opinion, are Burry's bets on US Treasury bonds crashing.


While inside China, it is virtually impossible to work in any capacity, in any field, without using WeChat.

Outside China, it is the most reliable means of communication to people without access to unrestricted internet. This has become even more difficult within the past year with the blocking of Shadowsocks proxies. You are unlikely to have friends with the ability or knowledge to install v2ray+vmess+websocket+tls for a reliable, unrestricted connection.

Your next best option is Facetime/iMessage, in my experience.


Does vmess still work? I haven’t been to China for 2 years but that’s what I relied on then to binge watch youtube


I can confirm that this method worked at least until early March of this year.


I am so happy Jitsi exists. My friends and I have a room that we regularly pop into to say hi or play games together.

The mobile app I downloaded through F-Droid works incredibly well, and for those of you Firefox users who aren't having the best experience, I recommend using the Electron desktop app [https://github.com/jitsi/jitsi-meet-electron/releases].

I've been using the Jitsi Electron app in conjunction with OBS + the VirtualCam plugin to share games, videos and my desktop. Hopefully I can convert more Zoom users.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: