How to build a Framework Desktop alternative that's actually repairable and upgradable and MORE powerful, for 35% less $$$
8mon 22d ago by piefed.social/u/artyom in buildapc
The Framework Desktop is a deep disappointment to me. Framework, the company that got into the business with an explicit purpose of building modular and repairable computers, went into a space where that was the norm (desktops), and introduced a PC that was none of those things, at an exorbitant price. When they debuted it, it was marketed specifically as a gaming machine. As much as I want to support them, I cannot reward them for this specific product, as it abandons their fundamental tenets.
Here's the build. You can see similar builds featured on many YT channels at this point with the new NV10 case and 5060 LP GPU.
Here's one from ETA Prime
And another from "MRGUI on PC"
This build: ~$1100
Comparable Framework build: ~$1700
I will concede the Framework is still better at a few things:
- Efficiency (I'm not sure that this is to any degree that's worth being factored in)
- Being that it's more efficient, it's also quieter
- Local LLMs (which no one should care about or be using)
- A bit thinner due to not have a dGPU
Local LLMs (which no one should care about or be using)
Yeah local LLMs are genuinely the ONLY thing the framework desktop is interesting for. For gaming, it is cheaper, more repairable, more upgradable, and more performant to just build a PC normally.
Why should no one care about or be using local LLMs, though? What a strange thing to say.
Because being blanket anti-llm is now the cool thing to do
Because they serve no purpose.
Is that why models on HuggingFace have monthly download counts in the millions?
No, that's because people are easily duped by unprecedented investments in marketing.
Sounds like you're coping. Especially since marketing for local LLMs is not a thing.
WTF is a cope? There's no way anyone actually thinks that marketing for LLMs is not a thing so I can only surmise that your intention is to argue in bad faith or otherwise troll. In which case, goodbye.
Could you provide a single example of marketing for a local LLM?
Do you have any evidence that investment in marketing of LLMs is actually "unprecedented" or trending any differently than marketing as a whole?
I can provide an example of a purpose for a local LLM: offline translation of documents written in a foreign language.
How much impact do you think marketing had on hugging face, considering it's free?
100% of it, considering HuggingFace wouldn't exist without LLM marketing.
It is typical anti-science conspiracy theorist "secret knowledge" mindset. People like OP want to feel like they know something others don't, so they allow themselves the delusion that, despite no personal expertise or knowledge of the topic, they actually know better than the companies risking billions on AI investments and expert researchers/academics with decades long careers in AI/ML.
They are the same as other science deniers, like anti-vaxxers believing they actually know better than doctors and medical researchers because of an anecodote they read online.
Admittedly, hugging face wasn't what it currently is in 2016, but the wiki suggests that it changed prior to 2021. Point still stands, as that's before LLMs even showed up in marketing.
That said, they certainly are painting with a broad brush.
Wow you're REALLY buttchugging the corporate koolaid...
claiming that working with free, open source, self hosted, privacy-focused local LLMs means you've drunk the corporate Kool aid is only further discrediting yourself on this topic...
Right, because FOSS LLMs couldn't possibly be influenced by the non-FOSS ones. That's stupid. That wouldn't make any sense at all...
because FOSS LLMs couldn't possibly be influenced by the non-FOSS ones.
Okay, so I have a question for you. If we replace "LLMs" with operating systems in that statement, you'd use the same logic to argue that it's drinking the corporate Kool aid to use linux, because it's been influenced by non-FOSS operating systems.
Do you see how embarrassing of an argument that is to try and make?
Let's shine some light on the elephant in the room. How much experience do you actually have with self-hosted LLMs? It's sounding like 0, in which case I'd speak a lot less authoritatively if that were the case.
It would be embarrassing, if that's what I had said. But Linux and other operating systems actually serve a purpose. They're not just bullshit machines.
Before we continue the conversation, I'm going to ask you again, how much experience do you actually have with local LLMs? I'm going to assume you just haven't found a use for it, and for some reason you have the gall to think that applies to everyone else, without ever touching it yourself.
I'll provide you a use-case I like. You can connect video feeds from self-hosted FOSS frigate NVR to a locally hosted LLM to provide searchable descriptions of tagged events, for example, you can search your events for "parked blue van", which was tagged by a local LLM, instead of trying to look through hundreds of events yourself, and without sending your private video feeds to any corporation.
Boom, local AI serves a purpose.
Boom, your earlier argument now reads as "using Linux is drinking the corporate Kool aid"
Are we done here?
I'm going to ask you again, how much experience do you actually have with local LLMs\
I'm not answering because I don't keep a log of time spent with LLMs. I can tell you it's many hours.
I'm going to assume you just haven't found a use for it, and for some reason you have the gall to think that applies to everyone else, without ever touching it yourself.
Well you know what happens when you assume. I haven't found a use for it because there aren't any, other than writing shitty social media posts.
You can connect video feeds from self-hosted FOSS frigate NVR to a locally hosted LLM to provide searchable descriptions of tagged events, for example, you can search your events for "parked blue van", which was tagged by a local LLM
LOOOOOLOL I see the problem here. You think LLM = AI. That's nothing to do with an LLM, that's just object detection/recognition.
Are we done here?
Yes, I'm glad I was able to clear that up for you!
LOOOOOLOL I see the problem here. You think LLM = AI. That's nothing to do with an LLM, that's just object detection/recognition.
....it's literally using ollama to produce textual descriptions of scenes. Maybe my example was too basic. As per the frigate documentation, you can search for "red sedan driving down a residential street on a sunny day". Object detection will not provide that context. That's why I added "parked" to blue car - common object detection algorithms like YOLO will only say "blue car" without any context of the scene it belongs in. Hell, I think YOLO itself will only say "car", IIRC.
Frigate does not provide that capability without LLMs
The only thing Ollama is doing is parsing your natural language into search queries for tags like "red car", "sedan", "residential street" etc. You could just as easily do the same thing by just typing in those tags. Like I said, pointless. And probably inaccurate.
The only thing Ollama is doing is parsing your natural language into search queries for tags like "red car", "sedan", "residential street" etc.
No, ollama is using models to provide further context on thumbnails containing tracked objects.. In textual form, that can be searched via semantic embeddings later on.
It blows my mind that you're trying to speak so authoritatively on something you've clearly misunderstood. I'm going to guess you haven't worked with frigate before?
Are you trying to understand me, or are you just defending yourself at all costs?
Please stop using your alt accounts to continue to harass me after I've blocked you. I told you I'm not interested in any bad faith discussions or personal insults.
"local LLMs have no use!"
"Here's a use"
"No, it's providing no utility in your example, because it's doing this."
"That's not even how the LLM is being used in my example"
"Please stop harassing me with your alt account"
Okay buddy. Turns out multiple people disagree with your uninformed takes. But I guess for some people it's easier to think a 2 year old, regularly active account is an alt than it is to acknowledge you were wrong lol.
EDIT: hold the fucking phone. Even if you think I was an alt, are you considering me providing a well documented counterexample to your assertion to be harassment? My guy, if you aren't informed about something it's okay to admit it. Everyone's wrong about a lot of things, what sets people apart is how they react when shown they were wrong. Life will be easier if you ask more questions when talking to strangers.
And you REALLY have totally outsmarted top researchers in AI/ML. Congratulations on knowing better! You should feel special!
Which corporation released local LLMs like OLMo? Oh wait that's right, it was a non-profit that released OLMo, not a corporation.
I didn't outsmart anyone except the stupid people being duped by trillions of dollars in marketing investment from the largest companies on the planet.
So let me understand how the world works, according to you:
Corporations have spent "trillions" of dollars to convince "stupid people" to download and run local LLMs... for free...
Actual facts: OpenAI, the largest player in AI, has raised a total of $59.7 billion in funding over the last 10 years. ... "trillions" 🤣
Wrong. I didn't say anything like that. You are once again intentionally trying to redirect the conversation back to local LLMs when I have done nothing but speak about LLMs in general. More marketing in proprietary LLMs = more adoption in LLMs, local or otherwise. You obviously know this, but once again have no interest in any good faith discussion, so this conversation is pointless.
I have done nothing but speak about LLMs in general
...
Local LLMs (which no one should care about or be using)
Because they serve no purpose.
^ this you? Lol.
You want to talk about good faith discussion, as if you're not literally making things up like "trillions" in "marketing investment" with no evidence or sources. As if you don't ignore any facts that disrupt your conspiracy theory, like OpenAI has only raised $59.7 billion, and cherry-pick the parts of the discussion that you respond to and the parts you ignore.
Who, exactly, is spending "trillions" of dollars on AI "marketing"? This question requires an answer for your claims to be taken seriously.
Are the companies that are spending billions to develop LLMs the ones pushing the "marketing" or are they the ones being duped by the "marketing"? Or somehow its both at once and they've tricked themselves into developing LLMs? Cognitive dissonance.
I saw it more as an AI bro machine rather than gaming-centric. It happens to play games fairly decently though provisions a great deal of gfx die area in doing so.
I agree with you that it's daft for a company like framework to put that out. wouldn't it be lovely if something like strix halo could work on a socketed system, perhaps leveraging a different memory technology than DDR5 UDIMMs
The whole point of this chip is unified memory, which can only be done using soldered memory. Using DIMMs would mean it must be typical shared memory, like any other APU/iGPU.
If the framework desktop came with a "normal" CPU/APU with non-soldered RAM, it would just be exactly the same as any other prebuilt desktop.
I'm aware of that and even though I was alluding to 'other memory technologies', I don't think something like CAMM2 (for example) could work for several reasons.
Then maybe they just shouldn't have made a desktop? It's like I said, this is a market segment that's already extremely repairable/upgradeable, and doesn't require their attention.
Not that I disagree with the sentiment here, but I wouldn't call the pricing exorbitant. Saving a third on something by building it yourself is a pretty standard rule of thumb.
You have to build them both.
Well yes, of course. I suppose 'building' was the wrong word to use in this context. My point was moreso about ease.
People that don't know anything about compatibility or physical fitment of components find value in offerings like Framework's desktop. Sort of like the difference in buying a bookshelf from Ikea over buying mdf panels from a hardware store. The Ikea costs a bit more, but not having to measure anything during assembly is worth the price of admission to most.
The hobbyist woodworker might go the route of measuring and cutting panels themselves, but that doesn't make the flat pack option exorbitant.