Live data from Hacker News

Running local LLMs offline on a ten-hour flight

deploy.live

91–100 of 112 posts

Re: Running local LLMs offline on a ten-hour flight

#91

Earlier quoted context omitted.

min_p author here. min_p is strictly better than top_p and top_k. The big labs don't know shit about sampling, and give absolutely nuts recommendations like this. set min_p to like 0.3 and ignore top_p and top_k and you'll be fine. There's better samplers now like top N sigma, top-h, P-less decoding, etc, but they're often not available in your LLM inference engine (i.e. vLLM)

I’m wondering though, what does extra creativity in code generation actually look like? How is the creativity expressed in code? Does the LLM reach for Bubble Sort instead of Quicksort? Maybe it decides that sorting only the first 10 elements of an array is enough? Funny variable names? Cursing in comments?

In this case, we are not arguing that min_p is better for "creative code" (you really don't want high temperature anywhere near your code generation, despite the "turning up the heat" framing of our paper) - at least in my post claiming min_p is strictly better than top_p above.

We are instead arguing that min_p handles truncating tokens that are more likely to lead to degeneration/looping because it is partially distribution aware. Fully distribution aware samplers like the ones I mentioned above (i.e. P-less decoding) are strictly superior due to using the whole distribution to decide the truncation at every time step.

Code hallucinations, like many LLM hallucinations, can be seen as accumulation of small amounts of "sampling errors".

Re: Running local LLMs offline on a ten-hour flight

#92
post #63

MacBook cable: 94W delivered Return flight will test this with the correct cable. I expect at least 16% improvement against the 70W cap Some plane sockets cut out completely if you attempt to draw more then the limit rather then continuing to provide power at the limit.

I’ve long suspected as such. Is there a way to power limit what my macbook pulls from the outlet?

Re: Running local LLMs offline on a ten-hour flight

#93

Earlier quoted context omitted.

I'd probably choose the window seat myself, because while it is cramped, it is predictably so. When I sit in an aisle seat, it's not as cramped but I regularly get shoulder checked by passing people or beverage carts. What really makes me nervous if I'm in an economy seat is the seat in front of me. Depending on how the seat is designed, if the person suddenly reclines (or hell, just flexes the seat a bunch while mov…

That was the first thing I thought of when I saw the image. That's a very expensive computer that you risk destroying when the 300lb guy in front of you decides to lean back. The ergonomics of using a laptop on an economy-class tray table are not worth it. You're sitting there like a T-rex trying to make your arms as small as possible to tap on the keys. And the vertical viewing angle to your screen sometimes prevent…

You just learn how to type with your toes. It is an easy skill to master and then comes handy when you drive, or sit at important meetings (you can just keep coding away) and noone's the wiser.

Re: Running local LLMs offline on a ten-hour flight

#96

My 14in M3 Max turns the fans way up when running local agentic coding for me to be comfortable with using it in a public place.

I don't use my M3 Max in public, but running local models on it makes me uncomfortable because of fans too.

I also have a Dell laptop that spin up fans if I open a text editor, and it feels normal, but on Mac, I feel like fans spinning up is me somehow abusing it and shortening its lifespan.

Re: Running local LLMs offline on a ten-hour flight

#97
post #70

Earlier quoted context omitted.

Running it on a Macbook Pro M5 48GB: -hf unsloth/Qwen3.6-27B-GGUF:UD-Q6_K_XL \ -c 128000 \ --parallel 1 \ --flash-attn on \ --no-context-shift \ --cache-type-k q8_0 \ --cache-type-v q8_0 \ --temp 0.6 \ --top-p 0.95 \ --top-k 20 \ --min-p 0.0 \ --presence_penalty 0.0 \ --reasoning on \ --jinja \ --chat-template-kwargs "{\"preserve_thinking\": true}" \ --spec-type ngram-simple \ --draft-max 64 \ --timeout 1800 Maybe so…

https://github.com/jundot/omlx note: 27b is going to be slow; use the 35b MoE if you want decent token/sec speed.

Many of us tested 27B and 35B side by side, and the dense model is significantly smarter. It indeed is slower, but 35B makes a lot of mistakes 27B doesn't.

Re: Running local LLMs offline on a ten-hour flight

#98

Earlier quoted context omitted.

I have been testing and using Qwen3.6 27B (running from my 3090) since it dropped and I genuinely think this is the first consumer hardware-grade model that can actually replace frontiers for a lot of workloads. I ran 8 tests on a variety of open-weights models, and opus 4.7 (1mil ctx version) and the little dense model was right behind it: https://github.com/sleepyeldrazi/llm_programming_tests/tree/... Of note is th…

> I have been testing With local models which are often benchmaxxed, testing unfortunately isn’t as predictive as you’d like.

I specifically tested on tasks I designed because I know every modern model, not only local ones, are bechmaxxed. The common benchmarks most labs use are (very likely) in their datasets to a degree (I'm assuming unintentionally, but is still highly probable) and there was a recent report on how easy it is to actually cheat them, as shown by people at UC Berkeley https://rdi.berkeley.edu/blog/trustworthy-benchmarks-cont/

That is precisely why my testing has been daily driving the model for everything + 8 tasks in a domain I care about. Could there be something very similar in their datasets? Of course, at least for most of the tasks, but if that lead to the good performance experience and results I'm getting, I am personally ok with that. I don't care how high the numbers are on the common benchmarks, only if it works well enough for me.

And if this model doesn't work for you, that's perfectly ok. Everyone has different needs from models. I was just impressed that it did for me, as it was a first from a local model.

Re: Running local LLMs offline on a ten-hour flight

#99
post #96

My 14in M3 Max turns the fans way up when running local agentic coding for me to be comfortable with using it in a public place.

I don't use my M3 Max in public, but running local models on it makes me uncomfortable because of fans too. I also have a Dell laptop that spin up fans if I open a text editor, and it feels normal, but on Mac, I feel like fans spinning up is me somehow abusing it and shortening its lifespan.

Reminds me of a previous generation (plastic shell) MBP drawing both AC and battery for sustained peak performance, fans spinning.

Re: Running local LLMs offline on a ten-hour flight

#100
post #57

Earlier quoted context omitted.

Obese people (250lb+) shouldn't even be allowed in Economy.

Sweet, free upgrades! Edit to be slightly less obtuse: surely you're not implying that a common carrier be allowed to discriminate based on facts about a passenger's body without making reasonable accommodations. Surely you're not implying that obese people not be allowed to fly at all. Surely you're not suggesting that fat people should just remove themselves from society so you don't have to deal with them. Therefo…

Free upgrades sound reasonable to me. Especially if obesity becomes less common via GLP-1s.
Post reply on HN