LM Studio
SoftwareWikipedia
50
MENTIONS
31
EPISODES
25
PODCASTS
Search complete. 50 mentions across 31 episodes found for "LM Studio".
Oct 16, 2026
Matt McDonaghHOST
so i think that you know super intelligence you know maybe it's not going to be some source of infinity above you but maybe it's going to be some source of coaching that you're going to use to get better at something so again like i know i sound like a little bit like a cheerleader for uh super intelligence in each of these episodes but i can't help but i'm so excited like this is this is, you can do anything you want with this.
Matt McDonaghHOST
You know, you can turn off the internet, you can pull in, you can first download the weights, of course, for the model that you want, but you could turn off the internet and you can, through LM Studio or another system, you can have a local conversation and you can be like, hi, my name is Becky.
Matt McDonaghHOST
Here's all the things I'm afraid of.
Matt McDonaghHOST
Here are all the things I want to do.
Stephen RoblesHOST
And what I realized...
Stephen RoblesHOST
So I downloaded LM Studio, and if you're not familiar, LM Studio is a free application that you can download that then has access to lots of open source models.
Jason AtenHOST
Open weight models.
Stephen RoblesHOST
Open weight models.
Jason AtenHOST
Yes, correct.
Stephen RoblesHOST
Like, z- zero cost because you download these models.
Stephen RoblesHOST
And so I will put a link to LM Studio.
Stephen RoblesHOST
They have two apps now.
Nathan HouseHOST
So you can try this tonight, actually.
Nathan HouseHOST
Install something called LM Studio, download the latest Llama model, and test it yourself.
Nathan HouseHOST
That's testable content you can practice in 10 minutes.
Nathan HouseHOST
The other big concept is RAG, retrieval-augmented generation.
Niall KitsonHOST
Sure.
Niall KitsonHOST
Well, if you happen to be using an app like LM Studio, you can download the current version of Mistral and give it a go on device.
Niall KitsonHOST
I'm not sharing your data with anyone.
Niall KitsonHOST
Yeah, maybe try it out.
Calvin Hendryx-ParkerHOST
So it detects your hardware.
Calvin Hendryx-ParkerHOST
It loads up a whole bunch of models from different providers like Ollama, Lama CPP, MLX, Docker Runner Models, LM Studios, and will let you just filter for what models would run best and gives it scores so you can actually see where it might be better fits.
Calvin Hendryx-ParkerHOST
I mean, be careful.
Calvin Hendryx-ParkerHOST
You're downloading potentially random models from potentially random places.
Federico ViticciHOST
They literally...
Federico ViticciHOST
Demos in person at WWDC, at product events, showing off the likes of LM Studio and Proplexity Computer operating local on-device agents with full disk access.
John VoorheesHOST
Yeah, and not only full disk access, but access across multiple mechs that are all matrixed together, right? Yes.
Federico ViticciHOST
Yes.
Herman PoppleberryHOST
There's a tracking issue, number three zero five nine nine, that states it plainly.
Herman PoppleberryHOST
SGLang's ROCm path is data center Instinct only, while every direct competitor already supports consumer cards, VLLM, Llama.cpp, Ollama, LM Studio, Cobalt CP.
CornHOST
So supports AMD and supports your AMD card are two different sentences.
Herman PoppleberryHOST
They're two different sentences, and the gap between them is where most of the disappointment lives.
Garry TrinderHOST
Um, so you know, I, I, I really liked that idea when we were talking about it, like we talked about Foundry Local, um, I think at one of the last Builds, about, you know, being able to run those local models.
Garry TrinderHOST
You got other options like Ollama, LM Studio, different places.
Garry TrinderHOST
But being able to control that, being able to say like, "Well, the execution is gonna happen in this one place, and I can define what goes in there, what goes out, what network, uh, it, it can access." You know, all those things, so you can really tie it down is, is where then I think we're gonna get a bit more of the, the trust around using these models.
Garry TrinderHOST
Because like you say, you, you hear lots of things in the news as well, and you know, sometimes that's people doing crazy things [laughs] and just-
6 MINS LATER
Kevin McDonnellHOST
You know, we're not gonna suddenly getting the average user.
Kevin McDonnellHOST
Thinking of someone that can be rude.
Kevin McDonnellHOST
Zoe, for example, wouldn't have to learn how to install, uh, Ollama or, uh, Windows ML or LM Studio.
Kevin McDonnellHOST
It will be built into the operating system itself and, and know when to use the right task for using that local model, which is, you know, that's what it should be really.
HackerNoon AINARRATOR
The active parameter count does not describe the full storage or memory footprint.
HackerNoon AINARRATOR
It runs in Llama CPP, LM Studio, Kobold CPP, and other GGUF-compatible runtimes.
HackerNoon AINARRATOR
The model card recommends NVFP4 or Apex quantization and Apex Compact for eight gigabytes and twelve gigabytes GPUs.
HackerNoon AINARRATOR
The listed license is Apache two zero, but check the terms of the upstream components before commercial deployment.
8 MINS LATER
HackerNoon AINARRATOR
Vision support in GGUF requires an file alongside the main GGUF.
HackerNoon AINARRATOR
Vocabulary and languages: two hundred and forty-eight K vocabulary, two hundred and one languages.
HackerNoon AINARRATOR
Model format and runtime: GGUF, compatible with Llama.cpp, LM Studio, KoboldCPP, and other GGUF-compatible runtimes.
HackerNoon AINARRATOR
Use Llama.cpp's flag for chat template handling.
Henrik OerdingHOST
Also man schließt dieses Teil an, man, äh, installiert da 'ne App.
Henrik OerdingHOST
Ich hab's jetzt mit LM Studio gemacht.
Henrik OerdingHOST
Das ist eine grafische Oberfläche für lokale KIs.
Henrik OerdingHOST
Das Programm schlägt dann direkt auch vor, welches Modell da passen würde.
21 more episodes mention LM Studio.
Create an account to see the whole feed, search across every transcript, and follow the entities you care about.