Hacker News Reader

At what temperature did you run it and what was your context limit?

Apparently there is a whole science behind running models. I have seen the instructions that unsloth publishes for their quants and depending on the model they'll tweak things like the temperature, top k, etc.

The size of the quantization you chose also makes a difference.

The GPU driver also plays an important role.

What was your approach? What software did you use to run the models?

mongrelionon Can I run AI locally?

What front-end framework did you use? I find the UI so visually appealing

mongrelionon Can I run AI locally?

Which quantization are you running and what context size? 32tok/s for that model on that card sounds pretty good to me!

mongrelionon Can I run AI locally?

It might be that the system prompt sent by codex is not optimal for that model. Try with open code and see if your results improve

mongrelionon How to run Qwen 3.5 locally

By anyone do you mean a well-established business or any entity willing to serve you?

mongrelionon Something is afoot in the land of Qwen

> [...] _but not necessarily use the right format._

This has also been my experience. But isn't the harness sending the instructions on how to invoke a tool? Maybe it is missing the formatting part. What do you think?

mongrelionon Ask HN: What Online LLM / Chat do you use?

Through my Kagi subscription I get access to quite a few models [1] but I tend to rely on Qwen3 (fast) for quick questions and Qwen3 (reasoning) when I want a more structured approach, for example, when I am researching a topic.

I have tried the same approach with Kimi K2.5 and GLM 5 but I keep going back fo Qwen3.

I also have access to Perplexity which is quite decent to be honest, but I prefer to keep everything in Kagi.

1: https://help.kagi.com/kagi/ai/assistant.html#available-llms

mongrelionon Right-sizes LLM models to your system's RAM, CPU, and GPU

Great idea of inferbench (similar to geekbench, etc.) but as of the time of writing, it's got only 83 submissions, which is underwhelming.

mongrelionon Right-sizes LLM models to your system's RAM, CPU, and GPU

> [...] it's much easier to fine-tune a "general" model into performing some very specific custom task (like classifying text, or translation, etc)

Is this fine-tunning process similar to training models? As in, do you need exhaustive resources? Or can this be done (realistically) on a consumer-grade GPU?

mongrelionon Right-sizes LLM models to your system's RAM, CPU, and GPU

> But are we really at the point yet where people are running local models without knowing what they are running them on..?

I can only speak for myself: it can be daunting for a beginner to figure out which model fits your GPU, as the model size in GB doesn't directly translate to your GPU's VRAM capacity.

There is value in learning what fits and runs on your system, but that's a different discussion.

mongrelionon Access to a Shared Unix Computer

Apparently there are a few more similar communities like the one from the post

https://tildeverse.org/members/

mongrelionon Pi – A minimal terminal coding harness

Pi ships with powerful defaults but skips features like sub-agents and plan mode

Does anyone have an idea as to why this would be a feature? don't you want to have a discussion with your agent to iron out the details before moving onto the implementation (build) phase?

In any case, looks cool :)

EDIT 1: Formatting EDIT 2: Thanks everyone for your input. I was not aware of the extensibility model that pi had in mind or that you can also iterate your plan on a PLAN.md file. Very interesting approach. I'll have a look and give it a go.

mongrelionon Alleged Distillation Attacks by DeepSeek, Moonshot AI, and MiniMax

I agree with you, especially with this:

They paid for the access the same as any other.

If anything, this makes them more legit than Anthropic because they are paying for the content, whereas Anthropic just stole *all* the data they got a hold of. So, in this case the Chinese AI labs stand on higher moral ground LOL.

mongrelionon Rise of the Triforce

The article touches a bit on how Sega basically lost. There is literally a whole documentary about this: Console Wars, where they go deep into how Sega lost the battle: https://en.wikipedia.org/wiki/Console_Wars_(film)

mongrelionon Show HN: Moltis – AI assistant with memory, tools, and self-extending skills

Hello. I am happy to take this for a spin.

I see that not all models available in my Github subscription are available (all models should be visible).

Further, is it possible to use openrouter with the current implementation? I couldn't figure it out by reading the documentation alone.

Thank you!

mongrelionon Improving 15 LLMs at Coding in One Afternoon. Only the Harness Changed

This is definitely a cool finding.

Have you investigated more on this topic? like, anything similar in concept that competes with Serena? if so, have you tested it/them? what are your thoughts?

mongrelionon LinkedIn checks for 2953 browser extensions

Curious question: why would they check for installed extensions on one's browser?

mongrelionon I replaced Windows with Linux and everything's going great

I had the dilemma of choosing between Bazzite and CachyOS or Manjaro for my workstation, which is also my gaming rig.

The whole immutable distro felt like a hinder for my workflow (running docker containers, etc.), so that's why I went for CachyOS.

mongrelionon I replaced Windows with Linux and everything's going great

I ran Archlinux as my main driver on both PC and Laptop for more than a decade but after having the opportunity to use a Windows machine with WSL and eventually WSL2, I felt like I had access to the best of both worlds: a Linux terminal for development (bash + tmux + vim, now bash + zellij + neovim) without the hassle of updates breaking things every few months and a out-of-the-box native gaming experience.

But with the enshitification of Windows (first all the spam and ads on the Start menu, then Microsoft forcing you to have an account to be able to use the machine and the expensive license for Windows Professional if you want access to Hyper-V, which I did), I did some research, tried a few new distros (Manjaro, Bazzite and CachyOS) and settled for CachyOS (gaming support was the main driver, based on Archlinux was secondary).

I do everything I did on Windows and some more: all the terminal stuff plus browsing, CAD modeling, 3D printing / slicing, Office stuff... I miss nothing. No more double partition to boot into Windows when I want to game.

My RX 9070 XT runs smoothly with no driver issues whatsoever. I even have tested the waters running some LLMs with LM Studio and that also worked out of the box.

The only thing that has been a bit meh are Teams and Slack and I believe that has to do with the fact that I ran them in Firefox. Once I ran Slack on Chromium, noise canceling was again available.

2009 was the year of Linux on desktop for me. 17 years later, after going back and forth between macOS and Windows, it feels good to be back home.

One last note in my random ramble is that I do not have as much spare time as before, and I had heard this from other people back in the day whenever I'd say I ran Archlinux on my machines, so I am going to repeat what others have said to me: it's really nice to not have to worry about much, be able to sit down and get productive right away. To me, CachyOS and KDE have made that idea my actual experience and for that I am grateful.

mongrelionon Ask HN: How can I get better at using AI for programming?

My experience with generating code with AI is very limited across a limited set of programming languages but whenever it has produced low quality code, it has been able to better itself with further instructions. Like "oh no, that is not the right naming convention. Please use instead" or "the choice of design pattern here is not great because ${reasons}. Produce 2 alternative solutions using x or y" and in nearly every case it produces satisfactory results.

Has this also been your experience?

mongrelionon Slax: Live Pocket Linux

This brings back good memories.

When I was in college I didn't have my own computer so I relied on the Windows machines from the library.

That's when I discovered Slax and Puppy Linux [1] and all its different variants. Such a beautiful thing to boot into these pendrive distros and discover the world of Linux through them without altering the host system.

Good times.

[1] https://puppylinux-woof-ce.github.io/

mongrelionon DeepSeek uses banned Nvidia chips for AI model, report says

I recommend everyone to watch GamersNexus' documentary on the NVIDIA AI GPU black market. They explain how companies like DeepSeek can get a hold of chips that are otherwise banned by the US government to export to China

https://www.youtube.com/watch?v=1H3xQaf7BFI

mongrelionon Ventoy: Create bootable USB drive for ISO/WIM/IMG/VHD(x)/EFI Files

I was going to ask how this would be better than any of the other options out there (like dd, the RPi imager and similar) but after seeing the README I consider this the superior alternative because you don't have to reflash the USB stick over and over again.

It supports multiple images at the same time, unlike the other solutions where one image take over the whole USB stick.

Love it.

I think your best bet is to traverse all the pipeline logs that make use of the action using Github's REST API.

It should be easy to do with thr Github CLI tool and some bash scripting.

Not sure how easy it'll be to parse the logs to look for a base64 string but it shouldn't be that complicated either.

mongrelionon Ask HN: Cheapest way to run a time-series database in cloud?

Is the cloud a must? What kind of performance are you after? You could repurpose an old desktop computer / server, slap a 4TB HDD on it and have it run at home or in your office.

The longer you use it, the cheaper it gets.

mongrelionon The Battery Revolution Is Finally Here

> how do we push all that energy into the battery at home in just 90 seconds?

If they are talking about an individual cell with a nominal voltage of 3.7V, then "any" modern fast charger will do.

I don't think they were talking about a full-blown EV power bank.

mongrelionon Show HN: OneUptime – open-source Datadog Alternative

Thanks for sharing this.

Others have mentioned Signoz which I have tried but for my homelab it just feels like too much.

OpenObserve might be the solution that I have been looking for.

mongrelionon Nintendo emulator 'Suyu' removed from Gitlab following DMCA request

Do you know if there is something similar for yuzu? did their code find a new home? forks are meh.

mongrelionon Swedish data brokers claim journalists' legal protection to evade EU law

Perhaps this is more security by obscurity but making it easier on criminals to find your personal information is not necessarily a good thing.

Imagine being a criminal and now you want to find your victim's personal information and you directly call Skatteverket...

Less compelling than going to hitta.se from a public, open wifi network