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

The gap between open weights LLMs and closed source LLMs

blog.doubleword.ai

Post reply on HN
Print view View original

231–240 of 264 posts

Re: The gap between open weights LLMs and closed source LLMs

#231

Post by port11 »

roenxi wrote: post #201
port11 wrote: post #190

Earlier quoted context omitted.

Cute. Climate change’s apocalyptical impact on food crops and cancer rates (post-ozone collapse) never convinced people to enact change. But hey, it’s open-model LLMs, the boogeyman! Can’t have that, it must be OpenAI or Anthropic safely controlling the market and calling all the shots.

Are you sure you haven't gotten your catastrophes crossed? Ozone depletion was a different crisis and people did enact change, the ozone hole has been closing fairly steadily. Wikipedia [0] thinks the prospects for the ozone layer are pretty good. [0] https://en.wikipedia.org/wiki/Ozone_depletion#Prospects_of_o...

Some climate models expect permafrost decay and/or mere GHG rise to erode the ozone layer. It’s not 1:1 proven, but there’s indication that in a post-4C world (with feedback loops), we might slowly lose our protection against the sun’s radiation.

I don’t recommend getting into the literature, it’s… depressing.

(Your point about the global, concerted effort to limit the “holes” in the ozone layer reminds me that we can, when pressed, come together to tackle serious issues.)

Re: The gap between open weights LLMs and closed source LLMs

#232

Post by ItsMonkk »

Smaug123 wrote: post #20
recursive wrote: post #12

Earlier quoted context omitted.

This seems backwards. Access to Fable can be removed. I don't see how an open weight model can ever be put back into the bag though.

The model itself, sure; the comment is about the production of more advanced models (to keep open weights near the frontier).

Governments can always offer prizes, and any model that sufficiently meets the criteria of the prize would win it and claim a large cash prize. Once claimed, the model would then be free to all forever.

Re: The gap between open weights LLMs and closed source LLMs

#233

Post by mirekrusin »

profsummergig wrote: post #5

IMHO, the biggest problem with the future of open weights models is that currently, open weights models are the result of philanthropy by some private org. (e.g. DeepSeek). The spigot can be turned off at any time. Until there's some sort of "community owned hardware", open weights models are always at risk of being discontinued.

Closed weight models are a result of philanthropy by some private investors.

Re: The gap between open weights LLMs and closed source LLMs

#234

Post by dabinat »

taffydavid wrote: post #178
dabinat wrote: post #33

I believe the open model party will eventually end. Perhaps because companies realize it’s too much of a commercial advantage, countries don’t want to give other countries commercial or military help, or maybe even an outright ban after someone uses an open model to guide them through how to make a bomb.

If we were going to ban technology because it helped people make bombs we wouldn't have access to much anymore.

Sure, but this administration doesn’t behave logically and big AI companies are already pushing the “AI is dangerous and only we should be trusted to wield it” angle.

Re: The gap between open weights LLMs and closed source LLMs

#235

Post by girvo »

Schiendelman wrote: post #219
girvo wrote: post #153

Earlier quoted context omitted.

Then you have no understanding of what DeepSeek has actually done. They publish their work openly: go have a look! Their architectural improvements are fascinating.

I think you're right - it was unclear to me , and now that I'm looking (especially at this morning's Deepseek news) I see they're doing quite a bit! Too bad I can't edit that previous comment to say I was wrong. :)

Hah the edit window here is the bane of my existence sometimes :)

They’re a super interesting company, and their innovations are fascinating

Re: The gap between open weights LLMs and closed source LLMs

#236

Post by trollbridge »

Gigachad wrote: post #100
verdverm wrote: post #99

Earlier quoted context omitted.

on the Will It Mythos benchmark, small models are punching way above their weight(s) gemma4-26B (#7) qwen-3.6-27B (#9) https://news.ycombinator.com/item?id=48640196

I've tried running qwen 3.6 locally and it felt like LLMs a year ago where you can get them to do some stuff but the tasks have to be very small and you have to course correct them a lot to the point it's hard to say it's any faster than doing it all yourself. Certainly the gap is closing but I feel it still makes more sense to pay pennies to run the full sized open models hosted on much better hardware.

What edition of Qwen 3.6? 35b-q6 (with MOE) has felt good enough for general purpose agentic coding.

You obviously need a 32GB card to run that (or a 64GB Mac), and realistically anything lesser than an RTX 5090 is going to be too slow for practical use.

Re: The gap between open weights LLMs and closed source LLMs

#237

Post by trollbridge »

sinuhe69 wrote: post #172

At this point, I think open weights vs proprietary models is a misnomer. First, we can not be sure the next release will remain open weights as Qwen 3.7 has showed. And second, they are all Chinese models. So instead of open weights, perhaps Chinese AI models is a better word choice.

It's not China's fault they're the only country releasing open weights.

Qwen has always alternated having an open release followed by a "max" release that isn't open weights.

Re: The gap between open weights LLMs and closed source LLMs

#238

Post by taffydavid »

hdgvhicv wrote: post #222
taffydavid wrote: post #208

Earlier quoted context omitted.

If only climate change was as easy a solve

Ozone action occurred before the rise of social media. We were fortunate. 20 years later and we’d have been screwed

I think it was more that the change required was relatively simple, and the industry was sort of already phasing out CFCs anyway. Banning them in fridges and aerosols was mostly frictionless and a win for everyone.

Turning the entire planet carbon neutral is so much harder a problem.

Also the fossil fuel industries have been trying to bury research on global warming since long before social media existed

Re: The gap between open weights LLMs and closed source LLMs

#239

Post by matheusmoreira »

woctordho wrote: post #159
matheusmoreira wrote: post #158

Earlier quoted context omitted.

I wish we had some kind of distributed training capability... Like Folding@home, but for LLMs.

See the recent advance of DiLoCo at Nous Research and Prime Intellect.

Really interesting!! This gives me hope!

Re: The gap between open weights LLMs and closed source LLMs

#240

Post by try-working »

jamiedborin1 wrote: post #174
profsummergig wrote: post #5

IMHO, the biggest problem with the future of open weights models is that currently, open weights models are the result of philanthropy by some private org. (e.g. DeepSeek). The spigot can be turned off at any time. Until there's some sort of "community owned hardware", open weights models are always at risk of being discontinued.

I am the original author of the post - thanks for reading it! I think the future of open weights models will be similar to fabless chip design companies. There will be companies that can train models and they will licence those models to inference companies that manage the APIs. The inference companies need much less capital and the training companies dont need to divert resources from training to inference. Some of…

This is exactly right.
Post reply on HN