Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Qwen 3.8

twitter.com

Post reply on HN
Print view View original

731–740 of 793 posts

Re: Qwen 3.8

#731

Post by gchamonlive »

whstl wrote: post #546
taosx wrote: post #462

Earlier quoted context omitted.

Exactly, and even when humans make bad decisions there is some friction, I feel that llm's don't have/notice that friction, they just bulldoze without caring about anything else.

Yeah. IME, LLMs actually introduces a lot of friction when you want to improve the design. They are 100% biased towards the status quo.

Not my experience. It does depend on the scale of the project, but for tens of thousands LOC, it's really good at taking instructions for architectural redesign.

LLM will take stuff at face value. If you tell it it's good, it'll gladly abide. Similarly if say it's crap it'll so the complete opposite.

Re: Qwen 3.8

#732

Post by 93po »

Suzuran wrote: post #681
sdoering wrote: post #628

Earlier quoted context omitted.

> human rights and all that rubbish Denying human rights. Classic. Sorry - you instantly lost any respect I could have maybe had for your opinion.

You are not allowed to discuss human rights until your government stops its unprovoked war on Iran, issues a full repudiation of the policies that led to it, and makes full reparations to the entire rest of the world whose energy prices you have driven up for the benefit of a handful of wealthy inside traders. Edit: Parent poster wasn't American. I incorrectly assumed their nationality.

And stops solitary confinement as a routine thing in prisons, which is literal torture. And stops death penalty. And provides basic healthcare instead of letting millions die.

Re: Qwen 3.8

#733

Post by tracker1 »

cnhwl wrote: post #583

I’m a developer from China. So, is this what Hacker News is all about? Whenever a model comes from China, the comments section stops discussing its technical architecture, optimisation points or real-world performance, and instead starts going on about politics, human rights and all that rubbish? To be honest, we Chinese IT professionals possess a genuine geek spirit. That’s why you’re lagging behind in open-source c…

It's a western site in English where a bulk of users are in the US, which has an adversarial relationship to the US... where many working in IT aren't even allowed to consider Chinese services or products directly (even if what you're getting was half made in China or controlled spaces).

I can't even consider Chinese AI tools for my own use on my own time on my own hardware without potentially risking firing or worse. Yeah, it's cool on a technical level.. but the "open" part of this and K3, etc. is of limited use depending on how/where you can actually use it yourself. Trust is another issue entirely.

Re: Qwen 3.8

#734

Post by Daishiman »

kelnos wrote: post #721
jpleyden98 wrote: post #665

Earlier quoted context omitted.

Perhaps whenever a US firm, eg. Open AI, Anthropic, Gemini, Grok etc. release a model. Top comments should be criticising US torture camps in Guantanamo bay or their recent bombing of a girls school in Iran.

The difference is that Chinese companies and their government are much more intertwined than their American (or Western, in general) counterparts. It seems much more reasonable to look at actions a Chinese company takes through the lens of what their government wants, than it does to do the same for a Western company.

> The difference is that Chinese companies and their government are much more intertwined than their American (or Western, in general) counterparts.

A very naive take that is trivially refuted when you look at the history of military contractors and how lobbying works in the US.

Military contractors, like banks and other pieces of corporate infrastructure, are extensions of the US government. They are beholden to profits first but US government interests as a really close second.

Re: Qwen 3.8

#736

Post by villish »

somenameforme wrote: post #644
villish wrote: post #609

Earlier quoted context omitted.

> Most people aren't entirely convinced that AI will ever be "too valuable" in the first place. There will be a generational leap that makes current LLMs obsolete in every way. Let's say China gets there first, why would the CCP let it be released publicly? That is an insanely valuable advantage in everything from war fighting to economics and more. Humans for millennia have used technological advances to make better…

You're making an assumption that's debatable, but let's just assume it's correct. The reality is that China won't be the only to achieve such. In the same way that the US put arguably the best minds in the world towards developing the atom bomb, and within a few years the Soviets had theirs as well. Shortly thereafter they were popping up all over Europe. It's kind of the story of LLMs in general - once people saw it…

> You're making an assumption that's debatable, but let's just assume it's correct. The reality is that China won't be the only to achieve such. In the same way that the US put arguably the best minds in the world towards developing the atom bomb, and within a few years the Soviets had theirs as well.

I never said other countries won't gain access to the technology. The point is that the 1st country to achieve that leap will get to use it to gain advantages in literally every sector.

I get some people are hoping that AI research is for the benefit of all, and I do too, but that is wishful thinking.

Re: Qwen 3.8

#737

Post by simonw »

simonw wrote: post #525

(I can't draw a pelican for this one because Alibaba Cloud have flagged my email address and won't let me pay them for access. So I'm waiting for the open weights release, or for the new model to show up on OpenRouter.)

I managed to get myself a working Qwen Cloud account, here's a pelican. The reasoning trace is pretty fun - three details I liked were "Could add helmet? No." and "Maybe add small bell? no." and "Need maybe add small fish in basket? Not necessary."

https://tools.simonwillison.net/markdown-svg-renderer#url=ht... (scroll to bottom for pelican)

Re: Qwen 3.8

#738

Post by andriy_koval »

cnhwl wrote: post #583

I’m a developer from China. So, is this what Hacker News is all about? Whenever a model comes from China, the comments section stops discussing its technical architecture, optimisation points or real-world performance, and instead starts going on about politics, human rights and all that rubbish? To be honest, we Chinese IT professionals possess a genuine geek spirit. That’s why you’re lagging behind in open-source c…

> That’s why you’re lagging behind in open-source competitions—it’s got nothing to do with politics

there are plenty oss geeks in US, but they don't have access to similar compute, alegelly subsidized by Chinese government.

Re: Qwen 3.8

#739

Post by simonw »

cypressf wrote: post #696
simonw wrote: post #525

(I can't draw a pelican for this one because Alibaba Cloud have flagged my email address and won't let me pay them for access. So I'm waiting for the open weights release, or for the new model to show up on OpenRouter.)

Here's what qwen 3.8 max preview generated when prompted with "create an svg of a pelican riding a bicycle" https://cypressf.com/qwen38-pelican.svg Notably it added a cheeky fish in the beak.

What method did you use to generate that one?

I tried using their OpenAI-compatible API and got one that's a lot less fun (no animation) - https://tools.simonwillison.net/markdown-svg-renderer#url=ht...

Post reply on HN