Comment by irishcoffee
21 hours ago
It is absolutely mind blowing to see some of the responses here. Open source, run-your-own, pay for nothing, we’re-all-nerds-that-buy-the-hardware-anyways ethos seems basically dead.
I guess I’m getting old. I own two 16gb cards and I use them for models, for gpu-pasthru for gaming, 3d model rendering, etc. 14 year old me is mortified at this community.
1) Different people might optimize for different things. There are people calculating that expensive hardware plus cheap rentals means owning isn’t optimizing, but there are people making choices that fit your preferences too.
2) I think it’s important to recognize that one of the things models are good for is astroturfing, and any given conversation you see may be direct or secondary effects of that (among other marketing).
The problem is accessibility. GPUs have gotten expensive and in some cases hard to get. You also need the supporting hardware to use them.
I'd love to have multiple large VRAM GPUs but I can't justify the costs when I have plenty of other more important things to spend that money on.
Times are changing. The open-weight models have needed time to catch up, but they're finally at a point now where we can get almost frontier level capabilities for coding.
I just wish we had a way to actually benchmark them properly though. Still seems no one has solved the problem of software architecture, brittleness and bloat as the codebase grows. Models love to add stuff, but they rarely clean up as they go. In a perfect world they'd do both near equally as they're developing.
It would be nice if there was an "architecture quality" benchmark that distilled the essence of what it means to have a good architecture, but I suppose that's an open research question with a lot of variables? Like how is good architecture actually quantified and measured? Is there a mechanism that can be re-used across all codebases to clearly denote one that is good and one that is bad, or is it highly subjective and depend on the lens you're looking at it from? Is there a lot more to it than just "how much refactoring effort is required to extend this in the future?".
Surely this is something that has been well researched - yet I never really hear anything about it. Makes me wonder why.
> Surely this is something that has been well researched - yet I never really hear anything about it. Makes me wonder why.
Occam’s razor rings true here: where’s the money in it?
14 year old me is mortified at this community.
Same here. There has to be someplace like this that's managed to cultivate a better crowd, but I'll be darned if I can find it.
This place is probably the best you’ll find. I actually found hn from a site called meta-somethingOrOther, a long time ago, and that site is probably the closest I can think of.
Hn really is a bit of an echo chamber, not that diverse opinions aren’t expressed, but the folks that voice opinions that don’t align with a specific set of values aren’t very well received here.
I’ll also say that this place has shaped my values, including making me change my opinions on things I severely disagreed with at the time. I’ve also said a lot of shit on here I wish I could wipe out.
I'm fine with diverse opinions, as long as they're not too diverse... and yes, there is such a thing as "too diverse." If I were to barge into an Amish town meeting and harangue them about how they should be using Qwen 3.6 27B Q8 to plan their crop rotation schedule, I would soon find myself heading out of town facing south on a northbound mule. And that's OK.
I feel the same here when "hackers" defend copyright maximalism, try to rehabilitate the Luddites, and argue that the Federal government should aggressively regulate AI models. Basically exhibiting both proud ignorance of history and reckless disregard for the future, all in one breath.
There are so many other places for that. So very, very many. Why do they come here? I spend time in those places as well, but I generally STFU when I have nothing to contribute, or when my core values conflict with their community charter.