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

There is minimal downside to switching to open models

marble.onl

Post reply on HN
Print view View original

331–340 of 351 posts

Re: There is minimal downside to switching to open models

#331

Post by 59nadir »

knollimar wrote: post #328
59nadir wrote: post #305

Earlier quoted context omitted.

Anyone who heavily uses LLMs for their work is pretty obviously ill-equipped for their work, and likely getting even worse at it with time, yes. You can throw around as many "things people from the US worship because they make money" positions/industries as you want. Nevermind that you're "pretty sure". People who both know what LLMs are actually capable of without major issues, plus are already capable enough to do…

Is the majority of work not BS gruntwork? My job has lots of low value stuff that must be done before I get high value tasks, so LLMs let me speed through to that point. (I use LLMs to spam disposable scripts in a construction environment)

If your "disposable scripts" require large-scale code generation, they're probably not "disposable" or "scripts".

Re: There is minimal downside to switching to open models

#332

Post by knollimar »

59nadir wrote: post #331
knollimar wrote: post #328

Earlier quoted context omitted.

Is the majority of work not BS gruntwork? My job has lots of low value stuff that must be done before I get high value tasks, so LLMs let me speed through to that point. (I use LLMs to spam disposable scripts in a construction environment)

If your "disposable scripts" require large-scale code generation, they're probably not "disposable" or "scripts".

Nah I'm in the trivial camp. Just the quantity beats doing it by hand, but the APIs for drafting are a bit esoteric to learn.

Re: There is minimal downside to switching to open models

#333

Post by 0xDEAFBEAD »

NicuCalcea wrote: post #327
0xDEAFBEAD wrote: post #234

Earlier quoted context omitted.

Who controls the definition of "fascist"? In any case, practically speaking, censorship helped the rise of the Nazis: https://www.fire.org/news/blogs/eternally-radical-idea/would... You can see far-right parties surging across Europe. Speech restriction isn't just authoritarian, it's also counterproductive. As an American I am actually quite worried about Europe's far right. Those guys are very scary, and it's creepy…

In spaces and interactions I control, the definition is up to me.

"Goebbels was in favor of free speech for views he liked. So was Stalin. If you’re really in favor of free speech, then you’re in favor of freedom of speech for precisely the views you despise. Otherwise, you’re not in favor of free speech."

Noam Chomsky

Re: There is minimal downside to switching to open models

#334

Post by 59nadir »

knollimar wrote: post #332
59nadir wrote: post #331

Earlier quoted context omitted.

If your "disposable scripts" require large-scale code generation, they're probably not "disposable" or "scripts".

Nah I'm in the trivial camp. Just the quantity beats doing it by hand, but the APIs for drafting are a bit esoteric to learn.

I suspect we agree more than our comments make it seem. There are code generation use cases that make sense for LLMs to do, I'm already with you on that, and I suspect that your situation is one of them.

Re: There is minimal downside to switching to open models

#335

Post by moomoo11 »

handoflixue wrote: post #325
moomoo11 wrote: post #313

Earlier quoted context omitted.

it's 2026. my work machine is a macbook pro with 64gb ram. i would rather spend a few hundred or thousands of dollars a month to make way more, than waste time and still lose to people who are using the latest commercial models which are 3 months ahead of the open source. what are you even talking about?

I am jokingly agreeing with you: Bill Gates famously claimed 640 kilobytes of RAM was sufficient for everyone, which felt like a funny parallel to people telling you "open source models are sufficient for everyone" There's obviously a competitive advantage to having better tools, and I find it amusing/absurd that some people don't get that.

my apologies. lol i see now. thank you, i did not intend to be a joke killer lol.

Re: There is minimal downside to switching to open models

#336

Post by mopsi »

0xDEAFBEAD wrote: post #321
mopsi wrote: post #291

Earlier quoted context omitted.

> This is 101-level UN stuff. If Ukrainian diplomats were unaware that Russia can veto Security Council resolutions, that means they were totally incompetent. There are ways around it, if there's a will: https://en.wikipedia.org/wiki/United_Nations_General_Assembl... It is safe to say that the present lack of leadership from the US was not foreseen at the time. It was unimaginable that Russia would launch a major gro…

>There are ways around it, if there's a will: https://en.wikipedia.org/wiki/United_Nations_General_Assembl ... The reason you do things via the Security Council is so you don't go to war with a nation that has nukes. The General Assembly is besides the point really. >It is safe to say that the present lack of leadership from the US was not foreseen at the time. It was unimaginable that Russia would launch a major gro…

  >  I've had enough. Time for a Swiss foreign policy. George Washington was right: No foreign entanglements or peacetime alliances. We'll handle our continent, and you handle yours.
Cooperation with allies has lasted so long that privileges like US warplanes being able to fly through Europe or use European facilities have become the normal and expected state of affairs. The US president spent a year complaining that allies are useless freeloaders and that the US would be better off without them, but the moment the war with Iran started, he threatened allies like Spain with economic warfare for not allowing US warplanes to use Spanish airbases to attack Iran[1], and demanded that European and Asian navies secure cargo ships through the Strait of Hormuz[2].

The whole huffing and puffing about alliances and isolationism is just the US refusing to keep their side of the deal while demanding that partners continue serving the US interests. It's strange to observe how shell-shocked US representatives are when that performative isolationism is treated seriously by partners, and they actually start taking independent steps like investing in their weapons programmes (instead of buying American).[3] It's as if they were seeking emotional satisfaction from the performance and weren't ready for American influence to wane.

[1] https://www.bbc.com/news/articles/c8r1mzd8vygo

[2] https://www.nytimes.com/2026/03/16/world/middleeast/trump-st...

[3] https://www.euronews.com/my-europe/2026/05/27/canada-chooses...

Re: There is minimal downside to switching to open models

#337

Post by 0xDEAFBEAD »

mopsi wrote: post #336
0xDEAFBEAD wrote: post #321

Earlier quoted context omitted.

>There are ways around it, if there's a will: https://en.wikipedia.org/wiki/United_Nations_General_Assembl ... The reason you do things via the Security Council is so you don't go to war with a nation that has nukes. The General Assembly is besides the point really. >It is safe to say that the present lack of leadership from the US was not foreseen at the time. It was unimaginable that Russia would launch a major gro…

> I've had enough. Time for a Swiss foreign policy. George Washington was right: No foreign entanglements or peacetime alliances. We'll handle our continent, and you handle yours. Cooperation with allies has lasted so long that privileges like US warplanes being able to fly through Europe or use European facilities have become the normal and expected state of affairs. The US president spent a year complaining that al…

As an isolationist, I consider it a positive if the US can no longer go to war in the Middle East.

Re: There is minimal downside to switching to open models

#338

Post by ignoramous »

epolanski wrote: post #304

I unsubscribed from Anthropic and our (EU-based) team is moving to an "ai-server" running opencode + GLM 5.2 and DS4. There are several benefits: - we cut AI spending by thousands - there is one AI server and starting different sessions for each user, one memory/skills/etc and everybody is involved into reviewing what went wrong and why. Harness finally makes sense and pays off more. - we can trust that the models ar…

> moving to an "ai-server" running opencode

What's the configuration of this server? How many do you need to deploy for your team size / estimated usage, if you're comfortable sharing that?

Re: There is minimal downside to switching to open models

#339

Post by sho »

sho wrote: post #142
linzhangrun wrote: post #131

Earlier quoted context omitted.

deepseek-v4-pro, probably the representative cheap opensouce LLM, was released in 2026.4 One year before, what OAI had in hand was gpt-4.1 and gpt-o3. I think it is not very controversial to say that deepseek is stronger than them, at most you can point to some post-training problems, basically the instability you mentioned. Also I am not sure if it is because the people who are best at using AI -- the people making…

You may be right, and I certainly hope so! But the question was about whether the Chinese labs will have fable-equivalence in 1 year. I am by no means some kind of insider, but knowing the vaguest outlines of what went into Mythos, they just can't do it. The compute is not there. The Chinese engineers are incredible, but they're not literal magicians. Of course there could be something incredible to come out of left…

I'm going to leave my above comment for embarrassment/posterity, but since writing it I've driven GLM5.2 much more extensively and I take it back. 5.2 is surprisingly, even shockingly good. It seems MUCH better than when I tried it on day one, and I probably shouldn't have drawn solid conclusions from that as models often struggle a little on launch.

I take it all back, 5.2 is very much competitive with any Opus and Fable seems very much in reach if they can continue making these leaps. It seems ZAI has quite a lot more up their sleeve than I had surmised.

Re: There is minimal downside to switching to open models

#340

Post by ignoramous »

aitchnyu wrote: post #248
flexagoon wrote: post #182

Earlier quoted context omitted.

OpenCode Go is $10/month and the limits are much more generous than those or Codex

After all the articles calculating OpenAI and Anthropic giving heavily subsidizing their subscriptions, how does OpenCode Go manage to be even cheaper?

> how does OpenCode Go manage to be even cheaper?

OpenCode may have partnerships with the Chinese labs or inference providers for better wholesale rates. In addition, right now, OpenCode Go rates are 4x more expensive for some of the more popular models (compared to providers on OpenRouter [0]), like MiMo v2.5 Pro & DeepSeek v4 Pro.

[0] https://news.ycombinator.com/item?id=48626751

Post reply on HN